[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17859982#comment-17859982 ] lkgen commented on KAFKA-1194: -- In order to use Kafka 3.7.0 in Windows with kafka streams I have created a different patch attempt this patch attempt is based on pull request https://github.com/apache/kafka/pull/6329 with modifications for 3.7.0 some more changes are based on exceptions info to reread files only when needed after close the attached patch file kafka370fixwin.patch can be applied with git apply on top of tag 3.7.0 > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > SendFor1194StreamCrash.zip, Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, kafka370fixwin.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17859975#comment-17859975
]
lkgen commented on KAFKA-1194:
--
Hi, I have been trying to use the pull request code of
[https://github.com/apache/kafka/pull/12331] in Windows with Kafka stream with
a test program
the test program seems to cause the Kafka to crash
Used the Kafka source code from pull request branch, set the following
server.properties modifications to cause the crash to happen after a short time
{code:java}
#comment out
log.retention.hours=168
#and add instead
log.retention.ms=3
#change existing
log.retention.check.interval.ms to 5000
#and add
log.roll.ms = 2{code}
The test program I use for Kafka streams source code is attached in the file
SendFor1194StreamCrash.zip
you can compile it with mvn package
Then copy the jar kafkaStreamTest-1.0-SNAPSHOT-jar-with-dependencies.jar to the
windows machine with kafka and run
{code:java}
java -jar kafkaStreamTest-1.0-SNAPSHOT-jar-with-dependencies.jar
{code}
In my test environment the Kafka crashed after several minutes with the error
{code:java}
[2024-06-25 19:13:20,749] ERROR Failed to clean up log for
__transaction_state-48 in dir C:\tmp\kafka-logs due to IOException
(org.apache.kafka.storage.internals.log.LogDirFailureChannel)
java.nio.file.FileAlreadyExistsException:
C:\tmp\kafka-logs\__transaction_state-48\.log ->
C:\tmp\kafka-logs\__transaction_state-48\.log.deleted
at
java.base/sun.nio.fs.WindowsException.translateToIOException(WindowsException.java:87)
at
java.base/sun.nio.fs.WindowsException.rethrowAsIOException(WindowsException.java:103)
at java.base/sun.nio.fs.WindowsFileCopy.move(WindowsFileCopy.java:395)
at
java.base/sun.nio.fs.WindowsFileSystemProvider.move(WindowsFileSystemProvider.java:292)
at java.base/java.nio.file.Files.move(Files.java:1422)
at
org.apache.kafka.common.utils.Utils.atomicMoveWithFallback(Utils.java:951)
at
org.apache.kafka.common.record.FileRecords.renameTo(FileRecords.java:247)
{code}
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components: log
>Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0
> Environment: window
>Reporter: Tao Qin
>Priority: Critical
> Labels: features, patch, windows
> Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt,
> SendFor1194StreamCrash.zip, Untitled.jpg, image-2018-09-12-14-25-52-632.png,
> image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch,
> kafka-bombarder.7z, screenshot-1.png
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> We tested it in windows environment, and set the log.retention.hours to 24
> hours.
> # The minimum age of a log file to be eligible for deletion
> log.retention.hours=24
> After several days, the kafka broker still cannot delete the old log file.
> And we get the following exceptions:
> [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task
> 'kafka-log-retention' (kafka.utils.KafkaScheduler)
> kafka.common.KafkaStorageException: Failed to change the log file suffix from
> to .deleted for log segment 1516723
> at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249)
> at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638)
> at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at
> scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59)
> at scala.collection.immutable.List.foreach(List.scala:76)
> at kafka.log.Log.deleteOldSegments(Log.scala:418)
> at
> kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314)
> at
> scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743)
> at scala.collection.Iterator$class.foreach(Iterator.scala:772)
> at
> scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573)
> at scala.collection.IterableLike$class.foreach(IterableLike.scala:73)
> at
> scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615)
> at
> s
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17567880#comment-17567880 ] Martin Pelak commented on KAFKA-1194: - Hi, I have created [a pull request|https://github.com/apache/kafka/pull/12331]. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17496586#comment-17496586 ] Martin Pelak commented on KAFKA-1194: - Hi [~mzinal], thank you for the fix. I have been testing Kafka 2.8.1 with your fix since yesterday on several Windows server machines and it seems that it fixed all of my issues I have faced on Windows. I was wondering whether you have considered creating [a pull request|https://github.com/apache/kafka/pulls]. Thank you! > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian Jira (v8.20.1#820001)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17477227#comment-17477227 ] Maksim Zinal commented on KAFKA-1194: - Duplicating my comment from KAFKA-2170 here. I've created a quick patch which seems to fix the issue on my system and on the Kafka version I use (2.8.0). I believe that the reason of lock failures are related to using RandomAccessFile Java API to open the files in some cases, while in other cases FileChannel.open() is used instead. When opening files with RandomAccessFile under Windows, FILE_SHARE_DELETE flag is not set, which leads to "access denied" errors when trying to rename or delete the open files. FileChannel.open() sets the FILE_SHARE_DELETE by default, as I checked on JDK 8 and 11. Here's the link to the branch based on tag 2.8.0: [https://github.com/zinal/kafka/tree/2.8.0_KAFKA-1194|[https://github.com/zinal/kafka/tree/2.8.0_KAFKA-1194].] Here are the exact changes implemented: [https://github.com/zinal/kafka/compare/2.8.0...zinal:2.8.0_KAFKA-1194] (plus jcenter and grgit stuff needed to run the build). > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename th
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17082819#comment-17082819 ] Chen Jinniu commented on KAFKA-1194: hey guys, seems Windows has been abandoned by kafka :( install linux in hyper-v is an option. wsl2 doesn't work. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17077642#comment-17077642
]
Fernando Grijalba commented on KAFKA-1194:
--
I'm experiencing this issue in version 2.4.1.
{code:java}
[2020-04-05 17:49:18,878] ERROR Error while deleting segments for topic-name in
dir C:\Kafka\data\kafka-logs (kafka.server.LogDirFailureChannel)[2020-04-05
17:49:18,878] ERROR Error while deleting segments for topic-name in dir
C:\Kafka\data\kafka-logs
(kafka.server.LogDirFailureChannel)java.nio.file.FileSystemException:
C:\Kafka\data\kafka-logs\topic-name\.timeindex ->
C:\Kafka\data\kafka-logs\topic-name\.timeindex.deleted: The
process cannot access the file because it is being used by another process.
at
java.base/sun.nio.fs.WindowsException.translateToIOException(WindowsException.java:92)
at
java.base/sun.nio.fs.WindowsException.rethrowAsIOException(WindowsException.java:103)
at java.base/sun.nio.fs.WindowsFileCopy.move(WindowsFileCopy.java:395) at
java.base/sun.nio.fs.WindowsFileSystemProvider.move(WindowsFileSystemProvider.java:292)
at java.base/java.nio.file.Files.move(Files.java:1425) at
org.apache.kafka.common.utils.Utils.atomicMoveWithFallback(Utils.java:795) at
kafka.log.AbstractIndex.renameTo(AbstractIndex.scala:209) at
kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:497) at
kafka.log.Log.$anonfun$deleteSegmentFiles$1(Log.scala:2225) at
kafka.log.Log.$anonfun$deleteSegmentFiles$1$adapted(Log.scala:2225) at
scala.collection.immutable.List.foreach(List.scala:392) at
kafka.log.Log.deleteSegmentFiles(Log.scala:2225) at
kafka.log.Log.removeAndDeleteSegments(Log.scala:2210) at
kafka.log.Log.$anonfun$deleteSegments$2(Log.scala:1722) at
scala.runtime.java8.JFunction0$mcI$sp.apply(JFunction0$mcI$sp.java:23) at
kafka.log.Log.maybeHandleIOException(Log.scala:2335) at
kafka.log.Log.deleteSegments(Log.scala:1713) at
kafka.log.Log.deleteOldSegments(Log.scala:1708) at
kafka.log.Log.deleteRetentionMsBreachedSegments(Log.scala:1785) at
kafka.log.Log.deleteOldSegments(Log.scala:1775) at
kafka.log.LogManager.$anonfun$cleanupLogs$3(LogManager.scala:982) at
kafka.log.LogManager.$anonfun$cleanupLogs$3$adapted(LogManager.scala:979) at
scala.collection.immutable.List.foreach(List.scala:392) at
kafka.log.LogManager.cleanupLogs(LogManager.scala:979) at
kafka.log.LogManager.$anonfun$startup$2(LogManager.scala:403) at
kafka.utils.KafkaScheduler.$anonfun$schedule$2(KafkaScheduler.scala:116) at
kafka.utils.CoreUtils$$anon$1.run(CoreUtils.scala:65) at
java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
at java.base/java.util.concurrent.FutureTask.runAndReset(FutureTask.java:305)
at
java.base/java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:305)
at
java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
at
java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
at java.base/java.lang.Thread.run(Thread.java:830) Suppressed:
java.nio.file.FileSystemException:
C:\Kafka\data\kafka-logs\topic-name\.timeindex ->
C:\Kafka\data\kafka-logs\topic-name\.timeindex.deleted: The
process cannot access the file because it is being used by another process.
at
java.base/sun.nio.fs.WindowsException.translateToIOException(WindowsException.java:92)
at
java.base/sun.nio.fs.WindowsException.rethrowAsIOException(WindowsException.java:103)
at java.base/sun.nio.fs.WindowsFileCopy.move(WindowsFileCopy.java:309) at
java.base/sun.nio.fs.WindowsFileSystemProvider.move(WindowsFileSystemProvider.java:292)
at java.base/java.nio.file.Files.move(Files.java:1425) at
org.apache.kafka.common.utils.Utils.atomicMoveWithFallback(Utils.java:792) ...
27 more
{code}
Any updates on when a fix is expected? Any known work around for version 2.4.1?
Thank you,
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components: log
>Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0
> Environment: window
>Reporter: Tao Qin
>Priority: Critical
> Labels: features, patch, windows
> Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt,
> Untitled.jpg, image-2018-09-12-14-25-52-632.png,
> image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch,
> kafka-bombarder.7z, screenshot-1.png
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> We tested it in windows environment, and set the log.retention.hours to 2
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17064643#comment-17064643 ] Stefan Meiforth Gulbrandsen commented on KAFKA-1194: Any new status on this? > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17040878#comment-17040878
]
Swaroop Kumar Sahu commented on KAFKA-1194:
---
Hi [~tqin],
The issue is exist in the latest stable version 2.4.0 also.
Please update the Affects Versions.
Logs:
at kafka.cluster.Partition.delete(Partition.scala:470)
at kafka.server.ReplicaManager.stopReplica(ReplicaManager.scala:360)
at
kafka.server.ReplicaManager.$anonfun$stopReplicas$2(ReplicaManager.scala:404)
at scala.collection.immutable.HashSet.foreach(HashSet.scala:932)
at kafka.server.ReplicaManager.stopReplicas(ReplicaManager.scala:402)
at kafka.server.KafkaApis.handleStopReplicaRequest(KafkaApis.scala:235)
at kafka.server.KafkaApis.handle(KafkaApis.scala:131)
at kafka.server.KafkaRequestHandler.run(KafkaRequestHandler.scala:70)
at java.lang.Thread.run(Thread.java:748)
*{color:#de350b}Suppressed: java.nio.file.AccessDeniedException:
C:\kafka_2.13-2.4.0\data\kafka\second_topic-4 ->
C:\kafka_2.13-2.4.0\data\kafka\second_topic-4.2f5254be07e947f7b0d999fa29a384f3-delete{color}*
at sun.nio.fs.WindowsException.translateToIOException(WindowsException.java:83)
at sun.nio.fs.WindowsException.rethrowAsIOException(WindowsException.java:97)
at sun.nio.fs.WindowsFileCopy.move(WindowsFileCopy.java:301)
at
sun.nio.fs.WindowsFileSystemProvider.move(WindowsFileSystemProvider.java:287)
at java.nio.file.Files.move(Files.java:1395)
at org.apache.kafka.common.utils.Utils.atomicMoveWithFallback(Utils.java:792)
... 17 more
[2020-02-20 17:01:10,761] INFO [ReplicaFetcherManager on broker 0] Removed
fetcher for partitions Set(first_topic-0, first_topic-1, first_topic-2)
(kafka.server.ReplicaFetcherManager)
[2020-02-20 17:01:10,762] INFO [ReplicaAlterLogDirsManager on broker 0] Removed
fetcher for partitions Set(first_topic-0, first_topic-1, first_topic-2)
(kafka.server.ReplicaAlterLogDirsManager)
[2020-02-20 17:01:10,765] INFO [ReplicaManager broker=0] Broker 0 stopped
fetcher for partitions first_topic-0,first_topic-1,first_topic-2 and stopped
moving logs for partitions because they are in the failed log directory
C:\kafka_2.13-2.4.0\data\kafka. (kafka.server.ReplicaManager)
[2020-02-20 17:01:10,766] INFO Stopping serving logs in dir
C:\kafka_2.13-2.4.0\data\kafka (kafka.log.LogManager)
[2020-02-20 17:01:10,770] ERROR Shutdown broker because all log dirs in
C:\kafka_2.13-2.4.0\data\kafka have failed (kafka.log.LogManager)
C:\kafka_2.13-2.4.0>
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components: log
>Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0
> Environment: window
>Reporter: Tao Qin
>Priority: Critical
> Labels: features, patch, windows
> Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt,
> Untitled.jpg, image-2018-09-12-14-25-52-632.png,
> image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch,
> kafka-bombarder.7z, screenshot-1.png
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> We tested it in windows environment, and set the log.retention.hours to 24
> hours.
> # The minimum age of a log file to be eligible for deletion
> log.retention.hours=24
> After several days, the kafka broker still cannot delete the old log file.
> And we get the following exceptions:
> [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task
> 'kafka-log-retention' (kafka.utils.KafkaScheduler)
> kafka.common.KafkaStorageException: Failed to change the log file suffix from
> to .deleted for log segment 1516723
> at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249)
> at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638)
> at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at
> scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59)
> at scala.collection.immutable.List.foreach(List.scala:76)
> at kafka.log.Log.deleteOldSegments(Log.scala:418)
> at
> kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314)
> at
> scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743)
> at scala.collection.Iterator$class.foreac
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16877731#comment-16877731 ] ASF GitHub Bot commented on KAFKA-1194: --- manmedia commented on pull request #3838: [KAFKA-1194 Invokes unmap if on Windows OS] URL: https://github.com/apache/kafka/pull/3838 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIR
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16836374#comment-16836374 ] Raoul de Haard commented on KAFKA-1194: --- Any updates on when this issue is being merged? (see:[https://github.com/apache/kafka/pull/6329#issuecomment-490901117] ) > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16828979#comment-16828979 ] Raoul de Haard commented on KAFKA-1194: --- All tests for PR: [https://github.com/apache/kafka/pull/6329] are green. Any update on when this is going to be merged? > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16777914#comment-16777914 ] Robert Brautigam commented on KAFKA-1194: - We have to install Kafka on Windows for a client and this bug is blocking us at the moment (since Kafka crashes on first log deletion and then shortly after each restart). For this reason I submitted a PR that seems to solve this problem for us: PR: https://github.com/apache/kafka/pull/6329 > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16777911#comment-16777911 ] ASF GitHub Bot commented on KAFKA-1194: --- robertbraeutigam commented on pull request #6329: KAFKA-1194: Fix renaming open files on Windows URL: https://github.com/apache/kafka/pull/6329 Running Kafka on Windows results in errors relating to renaming files that are locked (still mapped) by running Kafka. This patch unmmap()s and mmap()s a file before and after the actual renameTo() respectively. Patch should have no effect on Linux. This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16777910#comment-16777910 ] ASF GitHub Bot commented on KAFKA-1194: --- robertbraeutigam commented on pull request #6329: KAFKA-1194: Fix renaming open files on Windows URL: https://github.com/apache/kafka/pull/6329 Running Kafka on Windows results in errors relating to renaming files that are locked (still mapped) by running Kafka. This patch unmmap()s and mmap()s a file before and after the actual renameTo() respectively. Patch should have no effect on Linux. This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16770418#comment-16770418 ] ASF GitHub Bot commented on KAFKA-1194: --- kobihikri commented on pull request #6281: *DO NOT MERGE* - An attempt at https://issues.apache.org/jira/browse/KAFKA-1194 URL: https://github.com/apache/kafka/pull/6281 This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16770364#comment-16770364 ] ASF GitHub Bot commented on KAFKA-1194: --- kobihikri commented on pull request #6281: *DO NOT MERGE* - An attempt at https://issues.apache.org/jira/browse/KAFKA-1194 URL: https://github.com/apache/kafka/pull/6281 For some reason, running the tests locally I am receiving different failed tests than on the build server. As I would like to address all failing tests on the build server, I need to trigger a build. This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16741847#comment-16741847 ] Ismael Juma commented on KAFKA-1194: Promising that Windows 10 seems to support posix semantics now: FILE_DISPOSITION_POSIX_SEMANTICS https://docs.microsoft.com/en-us/windows-hardware/drivers/ddi/content/ntddk/ns-ntddk-_file_disposition_information_ex > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16700533#comment-16700533 ] lkgen commented on KAFKA-1194: -- Hi, if you are interested on a more complete approach for the problem you may inspect commit [https://github.com/lkgendev/kafka/commit/a9f49f51cc782b59bd2c20f9275dbf43ed8cc90e] > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16699243#comment-16699243 ] Kobi Hikri commented on KAFKA-1194: --- [~manme...@gmail.com], [~stephane.maa...@gmail.com], For some reason, I see that the automatic build for my PR fails but there is no error message which states the failing test Id. Can one of you (or anyone else, for that matter) give me a pointer as to how to proceed? > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16698875#comment-16698875 ] Kobi Hikri commented on KAFKA-1194: --- I've created a new pull request: [https://github.com/apache/kafka/pull/5945 ] > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16698860#comment-16698860 ] ASF GitHub Bot commented on KAFKA-1194: --- kobihikri closed pull request #5933: DO NOT MERGE - KAFKA-1194 PARTIAL bug fix URL: https://github.com/apache/kafka/pull/5933 This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16698604#comment-16698604 ] Kobi Hikri commented on KAFKA-1194: --- Hi all, As I want to move forward with my bug fix, I was running the tests. I've noticed failures that seemed unrelated to the region I worked on. So - I took latest version of the trunk branch and ran the unit tests and noticed the following failures: !image-2018-11-26-10-18-59-381.png! Am I supposed to just ignore that? if not, what is the next expected step? Best regards, Kobi. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, > image-2018-11-26-10-18-59-381.png, kafka-1194-v1.patch, kafka-1194-v2.patch, > kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16697135#comment-16697135 ] LiorK commented on KAFKA-1194: -- I have some code that appears to solve the problem on Windows, including the compaction, but it does not pass the Linux unit tests and may break some functionality Can you try using this code and help fix the unit tests ? is there a way to share code that does not pass unit testing > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16693461#comment-16693461 ] ASF GitHub Bot commented on KAFKA-1194: --- kobihikri opened a new pull request #5933: DO NOT MERGE - KAFKA-1194 PARTIAL bug fix URL: https://github.com/apache/kafka/pull/5933 This is a partial fix to the retention mechanism on Windows. It makes sure to close a segment file handles prior to deletion trial. Why is it partial? As it cannot work together will the log compaction policy. Should work with a similar to the following configuration: log.segment.bytes=1048576 log.retention.bytes= 5485760 log.retention.minutes = 1 log.retention.check.interval.minutes = 1 log.cleanup.policy = delete log.cleaner.enable = false # log.roll.ms = 1 <= Can be specified but will not cause any failure without the above two lines of configuration (log.cleanup.policy, log.cleaner.enable) ### Committer Checklist (excluded from commit message) - [ ] Verify design and implementation - [ ] Verify test coverage and CI build status - [ ] Verify documentation (including upgrade notes) This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(Sc
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16693465#comment-16693465 ] Kobi Hikri commented on KAFKA-1194: --- Hi all, In the above pull request, you will find a *partial fix* to the windows retention mechanism. Given a certain configuration, it is a workaround which allows using Kafka's retention mechanism on Windows. Please note - this fix *WILL NOT* work with *log.cleanup.policy = compact* Please share your thoughts. I believe with this fix, we can *close the critical bug* (as it has a workaround), and *open a lower severity bug* for utilizing the *log compaction* mechanism on Windows. Any comments are welcome, Kobi. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent b
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16688177#comment-16688177 ] Kobi Hikri commented on KAFKA-1194: --- Absolutely! > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16688174#comment-16688174 ] Kobi Hikri commented on KAFKA-1194: --- [~stephane.maa...@gmail.com], I was too quick to draw. The fix I've implemented does seem to help but it leads me to a similar problem with managing the consumer offsets files. Let me try and consolidate a fix with a solution for both problems. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16688175#comment-16688175 ] Stephane Maarek commented on KAFKA-1194: That's okay Here's another test case you want to test before pushing 1) start zookeeper 2) start kafka 3) create topic 4) push a few messages to it 5) delete topic > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16688085#comment-16688085 ] Kobi Hikri commented on KAFKA-1194: --- [~stephane.maa...@gmail.com], here it is: [https://github.com/apache/kafka/pull/5917 T|https://github.com/apache/kafka/pull/5917]hanks in advance, Kobi. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16688060#comment-16688060 ] Kobi Hikri commented on KAFKA-1194: --- [~stephane.maa...@gmail.com], Thanks. Doing it now and will update. [~manme...@gmail.com] - yes. However, unfortunately, this is not an option for a specific client. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16688044#comment-16688044 ] Stephane Maarek commented on KAFKA-1194: [~Kobi Hikri] excellent, I'll be happy to help out and test your PR. Here's the process: 1) go to github.com/apache/kafka 2) create a fork (top right corner). I'll assume your username is "user" 3) you now have a copy of the repo at github.com/user/kafka 4) add a remote to your git: git remote add myfork g...@gitlab.com:user/kafka.git 5) push your branch to your remote: git push -u myfork mybranch 6) go back to the web interface at github.com/apache/kafka 7) you should see a yellow bar to help you do a pull request. Else follow this guide: https://help.github.com/articles/creating-a-pull-request/ Hope that helps! Stephane > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a force
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16687995#comment-16687995 ] Kobi Hikri commented on KAFKA-1194: --- Thanks [~haraldk], I actually do understand the principle here :). I've prepared a fix for the issue and have stress tested it on Windows. However - I am unable to push the fix branch I've created in order to perform a PR. Can anyone assist with this? > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16688007#comment-16688007 ] Harald Kirsch commented on KAFKA-1194: -- [~Kobi Hikri] Good you know about! :) But I am not in a position to help with the PR, sorry. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16688008#comment-16688008 ] M. Manna commented on KAFKA-1194: - Just out of curiosity, have you considered dockerising your Kafka service, or does it not work for your business? Or perhaps, moving to Linux server? > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16687984#comment-16687984 ] Harald Kirsch commented on KAFKA-1194: -- [~Kobi Hikri] please be sure to understand the likely underlying principle issue with the JVM/mmap/Windows by following the pointers in https://issues.apache.org/jira/browse/KAFKA-1194?focusedCommentId=16580774&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-16580774 > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16687874#comment-16687874 ] Kobi Hikri commented on KAFKA-1194: --- Hi guys, Anyone from the core team, who want to help me in bringing this issue to a "fixed" state? I am currently "reverse engineering" the code, in order to provide a fix. Best regards, Kobi. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16647932#comment-16647932 ] Mario Giammarco commented on KAFKA-1194: Hello, we would like to use kakfa in a big project for our customer but infortunately in some deployments with need to put a local kafka on windows and this bug is a showstopper for us. We can try to help you in testing. Thanks, Mario > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16647927#comment-16647927 ] Bianchini Andrea commented on KAFKA-1194: - Hi guys, I'm new in this forum, I would like to help solving this issue. My skype is bianchinia maybe we can try to fix this together. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16637926#comment-16637926 ] Kobi Hikri commented on KAFKA-1194: --- [~manme...@gmail.com], Unfortunately, this workaround isn't (always) valid. I thought of the same approach and implemented a simple service which scans for "deleted" files and then deletes them. Sometimes the files are still being "held" by a a broker - even though they were marked for deletion, which then causes the "real" deletion command of the external, self implemented service, to fail. I think the problem here is a bit hard to solve with a "hack". I am still exploring the issue on my free time, but some direction might be useful. If any of you guys, which is more familiar with the retention mechanism would like to jump on a pair programming session with me over Skype or google hangouts - maybe we can combine our minds and find a solution. Best regards, Kobi. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mappe
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16625887#comment-16625887 ] shadi commented on KAFKA-1194: -- Thanks for the feedback. I tried to delete some log files beyond certain date manually on a cluster of three nodes. I took down the first node, deleted log files, and took it back up. I then took down the second node, deleted some log files, however it was unable to bounce as i was hit by this error: _Replica xxx for partition xxx reset its fetch offset from x to current leader 1's latest offset xxx (kafka.server.ReplicaFetcherThread)_ _Error getting offset for partition x to broker 1 (kafka.server.ReplicaFetcherThread)_ _java.lang.IllegalStateException: Compaction for partition cannot be aborted and paused since it is in LogCleaningPaused state._ > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > F
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16625600#comment-16625600 ] M. Manna commented on KAFKA-1194: - The best thing to do is to check the following: 1) Set the default value of log.segment.bytes to something slightly smaller e.g. 20 MB, 2) Disable log.cleaner.enable (i.e. false) 3) Allow rollover of new log files. 4) Now manually delete the files using some scripts. This is still not something I would recommend, but if Windows Operation is a MUST (without using containers), you want to try this out. Regards, > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16621754#comment-16621754 ] shadi commented on KAFKA-1194: -- If this old issue is not going to be fixed, then there must be be a manual procedure that's safe to delete old log files. Please share the instructions. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16612028#comment-16612028 ] Kobi Hikri commented on KAFKA-1194: --- Thanks for your time [~stephane.maa...@gmail.com]. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt, > Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16612019#comment-16612019
]
Stephane Maarek commented on KAFKA-1194:
Thanks [~Kobi Hikri]! That's what I thought, it was a quick fix for the log
cleaner, but it breaks the delete feature. At least the brokers don't crash
anymore
To be honest I won't pursue more fix because of the lack of time, but I really
hope someone in the core team can pick this up.
To me, it seems we have pinpointed through the PR all the points of potential
failures for Windows.
As for the quality of the fix, I would say very low, I honestly don't
understand much to the Windows file system and the order of Kafka's internal
for log cleaning and deletion.
I'm really hoping this is enough research so that [~ijuma] or [~lindong] can
probably determine a better fix.
In a super ideal world, the function to fix is
{code}Utils.atomicMoveWithFallback(src, dst);{code}
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components: log
>Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0
> Environment: window
>Reporter: Tao Qin
>Priority: Critical
> Labels: features, patch, windows
> Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt,
> Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch,
> kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> We tested it in windows environment, and set the log.retention.hours to 24
> hours.
> # The minimum age of a log file to be eligible for deletion
> log.retention.hours=24
> After several days, the kafka broker still cannot delete the old log file.
> And we get the following exceptions:
> [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task
> 'kafka-log-retention' (kafka.utils.KafkaScheduler)
> kafka.common.KafkaStorageException: Failed to change the log file suffix from
> to .deleted for log segment 1516723
> at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249)
> at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638)
> at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at
> scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59)
> at scala.collection.immutable.List.foreach(List.scala:76)
> at kafka.log.Log.deleteOldSegments(Log.scala:418)
> at
> kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314)
> at
> scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743)
> at scala.collection.Iterator$class.foreach(Iterator.scala:772)
> at
> scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573)
> at scala.collection.IterableLike$class.foreach(IterableLike.scala:73)
> at
> scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615)
> at
> scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742)
> at kafka.log.LogManager.cleanupLogs(LogManager.scala:314)
> at
> kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143)
> at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
> at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
> at java.lang.Thread.run(Thread.java:724)
> I think this error happens because kafka tries to rename the log file when it
> is still opened. So we should close the file first before rename.
> The index file uses a special data structure, the MappedByteBuffer. Javadoc
> describes it as:
>
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16612012#comment-16612012
]
ASF GitHub Bot commented on KAFKA-1194:
---
simplesteph closed pull request #5604: DO-NOT-MERGE KAFKA-1194: Alternative fix
URL: https://github.com/apache/kafka/pull/5604
This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:
As this is a foreign pull request (from a fork), the diff is supplied
below (as it won't show otherwise due to GitHub magic):
diff --git a/core/src/main/scala/kafka/log/AbstractIndex.scala
b/core/src/main/scala/kafka/log/AbstractIndex.scala
index ec9d55f89ac..9f1be4ac41a 100644
--- a/core/src/main/scala/kafka/log/AbstractIndex.scala
+++ b/core/src/main/scala/kafka/log/AbstractIndex.scala
@@ -202,8 +202,26 @@ abstract class AbstractIndex[K, V](@volatile var file:
File, val baseOffset: Lon
* @throws IOException if rename fails
*/
def renameTo(f: File) {
-try Utils.atomicMoveWithFallback(file.toPath, f.toPath)
-finally file = f
+if (!OperatingSystem.IS_WINDOWS) {
+ try Utils.atomicMoveWithFallback(file.toPath, f.toPath)
+ finally file = f
+} else {
+ // we get the file's position
+ val position = if (this.mmap == null) 0 else this.mmap.position()
+ if (OperatingSystem.IS_WINDOWS && this.mmap != null)
+// this sets mmap = null
+safeForceUnmap()
+ try {
+Utils.atomicMoveWithFallback(file.toPath, f.toPath)
+// we re-initialize mmap
+val raf = new RandomAccessFile(f, "rw")
+val len = raf.length()
+this.mmap = raf.getChannel.map(FileChannel.MapMode.READ_WRITE, 0, len)
+this.mmap.position(position)
+CoreUtils.swallow(raf.close(), this)
+ }
+ finally file = f
+}
}
/**
@@ -228,7 +246,7 @@ abstract class AbstractIndex[K, V](@volatile var file:
File, val baseOffset: Lon
// However, in some cases it can pause application threads(STW) for a
long moment reading metadata from a physical disk.
// To prevent this, we forcefully cleanup memory mapping within proper
execution which never affects API responsiveness.
// See https://issues.apache.org/jira/browse/KAFKA-4614 for the details.
- safeForceUnmap()
+ if (mmap != null) safeForceUnmap()
}
Files.deleteIfExists(file.toPath)
}
@@ -251,6 +269,7 @@ abstract class AbstractIndex[K, V](@volatile var file:
File, val baseOffset: Lon
/** Close the index */
def close() {
trimToValidSize()
+closeHandler()
}
def closeHandler(): Unit = {
diff --git a/core/src/main/scala/kafka/log/LogManager.scala
b/core/src/main/scala/kafka/log/LogManager.scala
index 32203acde9a..45cacd637f3 100755
--- a/core/src/main/scala/kafka/log/LogManager.scala
+++ b/core/src/main/scala/kafka/log/LogManager.scala
@@ -830,6 +830,7 @@ class LogManager(logDirs: Seq[File],
cleaner.abortCleaning(topicPartition)
cleaner.updateCheckpoints(removedLog.dir.getParentFile)
}
+ removedLog.close()
removedLog.renameDir(Log.logDeleteDirName(topicPartition))
checkpointLogRecoveryOffsetsInDir(removedLog.dir.getParentFile)
checkpointLogStartOffsetsInDir(removedLog.dir.getParentFile)
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components: log
>Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0
> Environment: window
>Reporter: Tao Qin
>Priority: Critical
> Labels: features, patch, windows
> Attachments: KAFKA-1194.patch, RetentionExpiredWindows.txt,
> Untitled.jpg, image-2018-09-12-14-25-52-632.png, kafka-1194-v1.patch,
> kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> We tested it in windows environment, and set the log.retention.hours to 24
> hours.
> # The minimum age of a log file to be eligible for deletion
> log.retention.hours=24
> After several days, the kafka broker still cannot delete the old log file.
> And we get the following exceptions:
> [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task
> 'kafka-log-retention' (kafka.utils.KafkaScheduler)
> kafka.common.KafkaStorageException: Failed to ch
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16611916#comment-16611916 ] Kobi Hikri commented on KAFKA-1194: --- Hi [~stephane.maa...@gmail.com], Apologies for the late response. I built the trunk branch merged with your PR. Log attached below. TL;DR: 1. The broker *no longer shuts-down* 2. Files marked for deletion are still not deleted. Few are deleted, but not all. 3. Plenty of "ERROR" log lines (as expected :)). > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16604427#comment-16604427 ] Kobi Hikri commented on KAFKA-1194: --- BTW, Steps to reproduce are as follows: 1. Set in server.properties: log.segment.bytes=1048576 log.retention.bytes= 5485760 log.retention.check.interval.ms=5000 And then ... Simply "bombard" a topic with data. My test application is a simple SpringBoot application which sends 100 UUIDs to the topic. I'm attaching the application for anyone's usage. You will see the log files are generated in the data directory immediately. After several seconds the broker will notify the error and shut down. [^kafka-bombarder.7z] > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, kafka-bombarder.7z, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16604413#comment-16604413 ] Kobi Hikri commented on KAFKA-1194: --- Sure Stephane. Will update shortly. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16604408#comment-16604408 ] Stephane Maarek commented on KAFKA-1194: hi [~Kobi Hikri], can you please test one more time https://github.com/apache/kafka/pull/5603 ? I hadn't fixed the class that is giving you errors If you manage to trigger errors, can you please give me the steps to reproduce? BTW, you may see a lot of "ERROR" logs, that's intended, just so I can see the code path. Please past full log somewhere if you get another error, that'll help > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16604253#comment-16604253 ] Kobi Hikri commented on KAFKA-1194: --- Hi all, Just built and tested @simplestephs proposed solution on Windows 10. The problem persists ... Stack trace below: [2018-09-05 13:28:30,705] ERROR Uncaught exception in scheduled task 'kafka-log-retention' (kafka.utils.KafkaScheduler) org.apache.kafka.common.errors.KafkaStorageException: Error while deleting segments for bombardedTopic-0 in dir C:\tmp\kafka-logs Caused by: java.nio.file.FileSystemException: C:\tmp\kafka-logs\bombardedTopic-0\.log -> C:\tmp\kafka-logs\bombardedTopic-0\.log.deleted: The process cannot access the file because it is being used by another process. at sun.nio.fs.WindowsException.translateToIOException(WindowsException.java:86) at sun.nio.fs.WindowsException.rethrowAsIOException(WindowsException.java:97) at sun.nio.fs.WindowsFileCopy.move(WindowsFileCopy.java:387) at sun.nio.fs.WindowsFileSystemProvider.move(WindowsFileSystemProvider.java:287) at java.nio.file.Files.move(Files.java:1395) at org.apache.kafka.common.utils.Utils.atomicMoveWithFallback(Utils.java:786) at org.apache.kafka.common.record.FileRecords.renameTo(FileRecords.java:211) at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:488) at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:1751) at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:1738) at kafka.log.Log$$anonfun$deleteSegments$1$$anonfun$apply$mcI$sp$1.apply(Log.scala:1309) at kafka.log.Log$$anonfun$deleteSegments$1$$anonfun$apply$mcI$sp$1.apply(Log.scala:1309) at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59) at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:48) at kafka.log.Log$$anonfun$deleteSegments$1.apply$mcI$sp(Log.scala:1309) at kafka.log.Log$$anonfun$deleteSegments$1.apply(Log.scala:1300) at kafka.log.Log$$anonfun$deleteSegments$1.apply(Log.scala:1300) at kafka.log.Log.maybeHandleIOException(Log.scala:1837) at kafka.log.Log.deleteSegments(Log.scala:1300) at kafka.log.Log.deleteOldSegments(Log.scala:1295) at kafka.log.Log.deleteRetentionSizeBreachedSegments(Log.scala:1383) at kafka.log.Log.deleteOldSegments(Log.scala:1361) at kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:875) at kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:873) at scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:733) at scala.collection.immutable.List.foreach(List.scala:392) at scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:732) at kafka.log.LogManager.cleanupLogs(LogManager.scala:873) at kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:395) at kafka.utils.KafkaScheduler$$anonfun$1.apply$mcV$sp(KafkaScheduler.scala:114) at kafka.utils.CoreUtils$$anon$1.run(CoreUtils.scala:63) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) at java.lang.Thread.run(Thread.java:745) Suppressed: java.nio.file.FileSystemException: C:\tmp\kafka-logs\bombardedTopic-0\.log -> C:\tmp\kafka-logs\bombardedTopic-0\.log.deleted: The process cannot access the file because it is being used by another process. at sun.nio.fs.WindowsException.translateToIOException(WindowsException.java:86) at sun.nio.fs.WindowsException.rethrowAsIOException(WindowsException.java:97) at sun.nio.fs.WindowsFileCopy.move(WindowsFileCopy.java:301) at sun.nio.fs.WindowsFileSystemProvider.move(WindowsFileSystemProvider.java:287) at java.nio.file.Files.move(Files.java:1395) at org.apache.kafka.common.utils.Utils.atomicMoveWithFallback(Utils.java:783) ... 32 more > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.pat
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16602298#comment-16602298 ] ASF GitHub Bot commented on KAFKA-1194: --- simplesteph opened a new pull request #5604: KAFKA-1194: Alternative fix URL: https://github.com/apache/kafka/pull/5604 Fix of KAFKA-1194 brought down to the minimal amount of code change. Related to #5603 In testing ### Committer Checklist (excluded from commit message) - [ ] Verify design and implementation - [ ] Verify test coverage and CI build status - [ ] Verify documentation (including upgrade notes) This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-co
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16602005#comment-16602005
]
ASF GitHub Bot commented on KAFKA-1194:
---
simplesteph closed pull request #5603: KAFKA-1194: Fixes for Windows
URL: https://github.com/apache/kafka/pull/5603
This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:
As this is a foreign pull request (from a fork), the diff is supplied
below (as it won't show otherwise due to GitHub magic):
diff --git
a/clients/src/main/java/org/apache/kafka/common/record/FileRecords.java
b/clients/src/main/java/org/apache/kafka/common/record/FileRecords.java
index cebb5fa09d1..38d9519b2d8 100644
--- a/clients/src/main/java/org/apache/kafka/common/record/FileRecords.java
+++ b/clients/src/main/java/org/apache/kafka/common/record/FileRecords.java
@@ -438,16 +438,20 @@ private static FileChannel openChannel(File file,
int initFileSize,
boolean preallocate) throws
IOException {
if (mutable) {
-if (fileAlreadyExists || !preallocate) {
-return FileChannel.open(file.toPath(),
StandardOpenOption.CREATE, StandardOpenOption.READ,
-StandardOpenOption.WRITE);
+if (fileAlreadyExists) {
+return FileChannel.open(file.toPath(),
StandardOpenOption.WRITE, StandardOpenOption.READ);
} else {
-RandomAccessFile randomAccessFile = new RandomAccessFile(file,
"rw");
-randomAccessFile.setLength(initFileSize);
-return randomAccessFile.getChannel();
+if (preallocate) {
+try (RandomAccessFile randomAccessFile = new
RandomAccessFile(file, "rw")) {
+randomAccessFile.setLength(initFileSize);
+}
+return FileChannel.open(file.toPath(),
StandardOpenOption.WRITE, StandardOpenOption.READ);
+} else {
+return FileChannel.open(file.toPath(),
StandardOpenOption.WRITE, StandardOpenOption.READ, StandardOpenOption.CREATE);
+}
}
} else {
-return FileChannel.open(file.toPath());
+return FileChannel.open(file.toPath(), StandardOpenOption.READ);
}
}
diff --git a/core/src/main/scala/kafka/log/AbstractIndex.scala
b/core/src/main/scala/kafka/log/AbstractIndex.scala
index ec9d55f89ac..aff3d363972 100644
--- a/core/src/main/scala/kafka/log/AbstractIndex.scala
+++ b/core/src/main/scala/kafka/log/AbstractIndex.scala
@@ -202,8 +202,26 @@ abstract class AbstractIndex[K, V](@volatile var file:
File, val baseOffset: Lon
* @throws IOException if rename fails
*/
def renameTo(f: File) {
-try Utils.atomicMoveWithFallback(file.toPath, f.toPath)
-finally file = f
+maybeLock(lock) {
+ val position = if (this.mmap == null) 0 else this.mmap.position()
+ if (OperatingSystem.IS_WINDOWS && this.mmap != null)
+forceUnmap()
+ try {
+Utils.atomicMoveWithFallback(file.toPath, f.toPath)
+if (OperatingSystem.IS_WINDOWS) {
+ val raf = new RandomAccessFile(f, "rw")
+ try {
+val len = raf.length()
+this.mmap = raf.getChannel.map(FileChannel.MapMode.READ_WRITE, 0,
len)
+this.mmap.position(position)
+ } finally {
+CoreUtils.swallow(raf.close(), this)
+ }
+}
+ } finally {
+file = f
+ }
+}
}
/**
@@ -228,7 +246,10 @@ abstract class AbstractIndex[K, V](@volatile var file:
File, val baseOffset: Lon
// However, in some cases it can pause application threads(STW) for a
long moment reading metadata from a physical disk.
// To prevent this, we forcefully cleanup memory mapping within proper
execution which never affects API responsiveness.
// See https://issues.apache.org/jira/browse/KAFKA-4614 for the details.
- safeForceUnmap()
+ CoreUtils.swallow(forceUnmap(), this)
+ // Accessing unmapped mmap crashes JVM by SEGV.
+ // Accessing it after this method called sounds like a bug but for
safety, assign null and do not allow later access.
+ mmap = null
}
Files.deleteIfExists(file.toPath)
}
@@ -251,6 +272,10 @@ abstract class AbstractIndex[K, V](@volatile var file:
File, val baseOffset: Lon
/** Close the index */
def close() {
trimToValidSize()
+inLock(lock) {
+ if (OperatingSystem.IS_WINDOWS)
+forceUnmap()
+}
}
def closeHandler(): Unit = {
@@ -316,8 +341,10 @@ abstract class AbstractIndex[K, V](@volatile var file:
File, val baseOffset: Lon
* Forcefully free the buffer's mmap.
*/
protected[log] def forceUnmap() {
-
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16602006#comment-16602006 ] ASF GitHub Bot commented on KAFKA-1194: --- simplesteph opened a new pull request #5603: KAFKA-1194: Fixes for Windows URL: https://github.com/apache/kafka/pull/5603 Adapt changes from #4431 to existing codebase ### Committer Checklist (excluded from commit message) - [ ] Verify design and implementation - [ ] Verify test coverage and CI build status - [ ] Verify documentation (including upgrade notes) This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16601613#comment-16601613
]
Stephane Maarek commented on KAFKA-1194:
[~lindong] Not fixed as of
{code}
[2018-09-02 11:03:12,789] INFO Kafka version : 2.1.0-SNAPSHOT
(org.apache.kafka.common.utils.AppInfoParser)
[2018-09-02 11:03:12,789] INFO Kafka commitId : 7299e1836ba2
(org.apache.kafka.common.utils.AppInfoParser)
{code}
Here are the full steps to reproduce and trigger:
{code}
C:\kafka_2.11-2.1.0-SNAPSHOT>kafka-topics.bat --zookeeper 127.0.0.1:2181
--topic second_topic --create --partitions 3 --replication-factor 1
WARNING: Due to limitations in metric names, topics with a period ('.') or
underscore ('_') could collide. To avoid issues it is best to use either, but
not both.
Created topic "second_topic".
C:\kafka_2.11-2.1.0-SNAPSHOT>kafka-console-producer.bat --broker-list
127.0.0.1:9092 --topic second_topic
>hello
>world
>hello
>Terminate batch job (Y/N)? Y
C:\kafka_2.11-2.1.0-SNAPSHOT>kafka-topics.bat --zookeeper 127.0.0.1:2181
--topic second_topic --delete
Topic second_topic is marked for deletion.
Note: This will have no impact if delete.topic.enable is not set to true.
{code}
Which triggers Kafka shutdown with:
{code}
[2018-09-02 11:04:15,460] ERROR Error while renaming dir for second_topic-1 in
log dir C:\kafka_2.11-2.1.0-SNAPSHOT\data\kafka
(kafka.server.LogDirFailureChannel)
java.nio.file.AccessDeniedException:
C:\kafka_2.11-2.1.0-SNAPSHOT\data\kafka\second_topic-1 ->
C:\kafka_2.11-2.1.0-SNAPSHOT\data\kafka\second_topic-1.d1ceee24d7474152b6fedd61903449e5-delete
at sun.nio.fs.WindowsException.translateToIOException(Unknown Source)
at sun.nio.fs.WindowsException.rethrowAsIOException(Unknown Source)
at sun.nio.fs.WindowsFileCopy.move(Unknown Source)
at sun.nio.fs.WindowsFileSystemProvider.move(Unknown Source)
at java.nio.file.Files.move(Unknown Source)
at
org.apache.kafka.common.utils.Utils.atomicMoveWithFallback(Utils.java:809)
at kafka.log.Log$$anonfun$renameDir$1.apply$mcV$sp(Log.scala:689)
at kafka.log.Log$$anonfun$renameDir$1.apply(Log.scala:687)
at kafka.log.Log$$anonfun$renameDir$1.apply(Log.scala:687)
at kafka.log.Log.maybeHandleIOException(Log.scala:1842)
at kafka.log.Log.renameDir(Log.scala:687)
at kafka.log.LogManager.asyncDelete(LogManager.scala:833)
at kafka.cluster.Partition$$anonfun$delete$1.apply(Partition.scala:271)
at kafka.cluster.Partition$$anonfun$delete$1.apply(Partition.scala:265)
at kafka.utils.CoreUtils$.inLock(CoreUtils.scala:251)
at kafka.utils.CoreUtils$.inWriteLock(CoreUtils.scala:259)
at kafka.cluster.Partition.delete(Partition.scala:265)
at kafka.server.ReplicaManager.stopReplica(ReplicaManager.scala:340)
at
kafka.server.ReplicaManager$$anonfun$stopReplicas$2.apply(ReplicaManager.scala:370)
at
kafka.server.ReplicaManager$$anonfun$stopReplicas$2.apply(ReplicaManager.scala:368)
at scala.collection.Iterator$class.foreach(Iterator.scala:891)
at scala.collection.AbstractIterator.foreach(Iterator.scala:1334)
at scala.collection.IterableLike$class.foreach(IterableLike.scala:72)
at scala.collection.AbstractIterable.foreach(Iterable.scala:54)
at kafka.server.ReplicaManager.stopReplicas(ReplicaManager.scala:368)
at kafka.server.KafkaApis.handleStopReplicaRequest(KafkaApis.scala:200)
at kafka.server.KafkaApis.handle(KafkaApis.scala:111)
at kafka.server.KafkaRequestHandler.run(KafkaRequestHandler.scala:69)
at java.lang.Thread.run(Unknown Source)
Suppressed: java.nio.file.AccessDeniedException:
C:\kafka_2.11-2.1.0-SNAPSHOT\data\kafka\second_topic-1 ->
C:\kafka_2.11-2.1.0-SNAPSHOT\data\kafka\second_topic-1.d1ceee24d7474152b6fedd61903449e5-delete
at sun.nio.fs.WindowsException.translateToIOException(Unknown
Source)
at sun.nio.fs.WindowsException.rethrowAsIOException(Unknown
Source)
at sun.nio.fs.WindowsFileCopy.move(Unknown Source)
at sun.nio.fs.WindowsFileSystemProvider.move(Unknown Source)
at java.nio.file.Files.move(Unknown Source)
at
org.apache.kafka.common.utils.Utils.atomicMoveWithFallback(Utils.java:806)
... 23 more
[2018-09-02 11:04:15,460] INFO [ReplicaManager broker=0] Stopping serving
replicas in dir C:\kafka_2.11-2.1.0-SNAPSHOT\data\kafka
(kafka.server.ReplicaManager)
{code}
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components:
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16597263#comment-16597263 ] Kobi Hikri commented on KAFKA-1194: --- Thank you very much for your reply. I did read the ticket you mentioned and even went over the suggested fix in the latest PR :) I will try to reproduce the problem with the latest PR and update. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16597234#comment-16597234 ] M. Manna commented on KAFKA-1194: - [~Kobi Hikri] Please check what I have mentioned with the relevant ticket. Yes you will always find this problem as the fix hasn't gone in yet. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16597224#comment-16597224 ] Kobi Hikri commented on KAFKA-1194: --- I reproduced the problem yesterday, working with latest kafka version. It was reproduced on windows 10. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16597218#comment-16597218 ] M. Manna commented on KAFKA-1194: - [~Kobi Hikri] I believe [7278|https://issues.apache.org/jira/browse/KAFKA-7278] has already, and "Potentially" fixed the issue. [~lindong] You want to elaborate whether that's the case. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16597214#comment-16597214 ] Kobi Hikri commented on KAFKA-1194: --- Hi all, I would like to take a look at the bug and try and fix it. Anyone working on it currently? Wouldn't like to interfere ... > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16588446#comment-16588446 ] Harald Kirsch commented on KAFKA-1194: -- Most likely, and if the root cause is [https://bugs.openjdk.java.net/browse/JDK-4724038, i|https://bugs.openjdk.java.net/browse/JDK-4724038]t is an unfortunate interaction between the JVM not releasing a file handle of a mapped file immediately when it is closed and Window's locking of open files. So yes, there is a good chance you get the problem in a Windows container. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16588155#comment-16588155 ] Charith Fernando commented on KAFKA-1194: - Will this work If I run on a windows container? * In my local machine which is Windows 10 - the problem occurs * I tried with docker on windows 10 and a linux container - it works without a problem * I have to deploy the application on Windows server 2016 which makes me deploy this app on a windows container. * Therefore will the same problem occur on a windows container? > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16580836#comment-16580836 ] M. Manna commented on KAFKA-1194: - [~haraldk] not sure about this jdk bug The issue happens mainly on WindowsLinux works just fine and also in containers Recently [~lindong] has identified a possible bug where asyc delete shouldn't be called on certain segments. I will look forward to that patch > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16580774#comment-16580774 ] Harald Kirsch commented on KAFKA-1194: -- As far as I understand the issue, it might have to do with this JDK bug: [https://bugs.openjdk.java.net/browse/JDK-4724038] which says that once a file is mapped into memory, there is not really a way to unmap it. One has to hope for the finalize() method to run eventually. On Windows, the effect is that files are kept locked and cannot be deleted. The bug was opened in 2002 :), so without someone really knowledgeable digging into this, the situation will hardly improve :( The Lucene folks seem to be able to deal with this somehow, or they can just wait long enough for their mmapped index files to get unmapped. More background for anyone interested might be found via: https://issues.apache.org/jira/browse/LUCENE-6989 > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16580480#comment-16580480 ] Charith Fernando commented on KAFKA-1194: - This issue still exists on my environment. Seems like after the retention period it cannot clean up the logs. If I have to manually to perform the cleaning up operation this will be a problem for us to go into production. Is there any solution/patch for this? Why isn't this bug addressed for the last few years? > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16506062#comment-16506062
]
Michael Hartle commented on KAFKA-1194:
---
I can reproduce the crash on Windows with Apache Kafka 1.1.0 with
{quote}delete.topic.enable = true
log.cleaner.enable = false
{quote}
just by deleting a topic. Restarting the broker fails until the log files are
deleted manually. Disabling the Cleaner may work in other circumstances, but at
least not in my case when trying to delete a topic.
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components: log
>Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0
> Environment: window
>Reporter: Tao Qin
>Priority: Critical
> Labels: features, patch, windows
> Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch,
> kafka-1194-v2.patch, screenshot-1.png
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> We tested it in windows environment, and set the log.retention.hours to 24
> hours.
> # The minimum age of a log file to be eligible for deletion
> log.retention.hours=24
> After several days, the kafka broker still cannot delete the old log file.
> And we get the following exceptions:
> [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task
> 'kafka-log-retention' (kafka.utils.KafkaScheduler)
> kafka.common.KafkaStorageException: Failed to change the log file suffix from
> to .deleted for log segment 1516723
> at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249)
> at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638)
> at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at
> scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59)
> at scala.collection.immutable.List.foreach(List.scala:76)
> at kafka.log.Log.deleteOldSegments(Log.scala:418)
> at
> kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314)
> at
> scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743)
> at scala.collection.Iterator$class.foreach(Iterator.scala:772)
> at
> scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573)
> at scala.collection.IterableLike$class.foreach(IterableLike.scala:73)
> at
> scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615)
> at
> scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742)
> at kafka.log.LogManager.cleanupLogs(LogManager.scala:314)
> at
> kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143)
> at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
> at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
> at java.lang.Thread.run(Thread.java:724)
> I think this error happens because kafka tries to rename the log file when it
> is still opened. So we should close the file first before rename.
> The index file uses a special data structure, the MappedByteBuffer. Javadoc
> describes it as:
> A mapped byte buffer and the file mapping that it represents remain valid
> until the buffer itself is garbage-collected.
> Fortunately, I find a forceUnmap function in kafka code, and perhaps it can
> be used to free the MappedByteBuffer.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16480453#comment-16480453
]
Christoph Schmidt commented on KAFKA-1194:
--
(!) The only permanently functioning workaround currently known appears to be
to completely disable the Cleaner:
{quote} {{log.cleaner.enable = false}}
{quote}
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components: log
>Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0
> Environment: window
>Reporter: Tao Qin
>Priority: Critical
> Labels: features, patch, windows
> Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch,
> kafka-1194-v2.patch, screenshot-1.png
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> We tested it in windows environment, and set the log.retention.hours to 24
> hours.
> # The minimum age of a log file to be eligible for deletion
> log.retention.hours=24
> After several days, the kafka broker still cannot delete the old log file.
> And we get the following exceptions:
> [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task
> 'kafka-log-retention' (kafka.utils.KafkaScheduler)
> kafka.common.KafkaStorageException: Failed to change the log file suffix from
> to .deleted for log segment 1516723
> at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249)
> at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638)
> at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at
> scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59)
> at scala.collection.immutable.List.foreach(List.scala:76)
> at kafka.log.Log.deleteOldSegments(Log.scala:418)
> at
> kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314)
> at
> scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743)
> at scala.collection.Iterator$class.foreach(Iterator.scala:772)
> at
> scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573)
> at scala.collection.IterableLike$class.foreach(IterableLike.scala:73)
> at
> scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615)
> at
> scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742)
> at kafka.log.LogManager.cleanupLogs(LogManager.scala:314)
> at
> kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143)
> at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
> at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
> at java.lang.Thread.run(Thread.java:724)
> I think this error happens because kafka tries to rename the log file when it
> is still opened. So we should close the file first before rename.
> The index file uses a special data structure, the MappedByteBuffer. Javadoc
> describes it as:
> A mapped byte buffer and the file mapping that it represents remain valid
> until the buffer itself is garbage-collected.
> Fortunately, I find a forceUnmap function in kafka code, and perhaps it can
> be used to free the MappedByteBuffer.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16479130#comment-16479130 ] Grégory R. commented on KAFKA-1194: --- This issue is reproduced with kafka 1.1.0 > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16463443#comment-16463443
]
Christoph Schmidt commented on KAFKA-1194:
--
Would love to have this upgraded to Blocker criticality, since the issue
produces recurring crashes, and the proposed workarounds ("just use Linux") are
so fundamental that they're basically not applicable.
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components: log
>Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0
> Environment: window
>Reporter: Tao Qin
>Priority: Critical
> Labels: features, patch, windows
> Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch,
> kafka-1194-v2.patch, screenshot-1.png
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> We tested it in windows environment, and set the log.retention.hours to 24
> hours.
> # The minimum age of a log file to be eligible for deletion
> log.retention.hours=24
> After several days, the kafka broker still cannot delete the old log file.
> And we get the following exceptions:
> [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task
> 'kafka-log-retention' (kafka.utils.KafkaScheduler)
> kafka.common.KafkaStorageException: Failed to change the log file suffix from
> to .deleted for log segment 1516723
> at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249)
> at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638)
> at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at
> scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59)
> at scala.collection.immutable.List.foreach(List.scala:76)
> at kafka.log.Log.deleteOldSegments(Log.scala:418)
> at
> kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316)
> at
> kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314)
> at
> scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743)
> at scala.collection.Iterator$class.foreach(Iterator.scala:772)
> at
> scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573)
> at scala.collection.IterableLike$class.foreach(IterableLike.scala:73)
> at
> scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615)
> at
> scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742)
> at kafka.log.LogManager.cleanupLogs(LogManager.scala:314)
> at
> kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143)
> at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
> at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
> at java.lang.Thread.run(Thread.java:724)
> I think this error happens because kafka tries to rename the log file when it
> is still opened. So we should close the file first before rename.
> The index file uses a special data structure, the MappedByteBuffer. Javadoc
> describes it as:
> A mapped byte buffer and the file mapping that it represents remain valid
> until the buffer itself is garbage-collected.
> Fortunately, I find a forceUnmap function in kafka code, and perhaps it can
> be used to free the MappedByteBuffer.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16458582#comment-16458582 ] ASF GitHub Bot commented on KAFKA-1194: --- mmanna-sapfgl opened a new pull request #4947: KAFKA-1194 Forces unmapping of buffer if on Windows URL: https://github.com/apache/kafka/pull/4947 In response to KAFKA-1194 issue, it now forces buffer to unmap if running on Windows OS. Invidual Unit Test 'testChangeFileSuffixes' on LogSegmentTest.scala passes. but other file IO related unit tests still fail on Windows (Which is irrelevant to this code change). This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunate
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16458526#comment-16458526 ] M. Manna commented on KAFKA-1194: - I will adding a patch to the trunk and raise a PR. In the past I have raised a PR and all existing test should have just passed. There are some bizzare JDK compatibility tests which fails for unrelated reasons. May be someone from permanent members should review the PR. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16441993#comment-16441993 ] Christoph Schmidt commented on KAFKA-1194: -- Seeing this in a project which is supposed to collapse events almost immediately, with consumers only ever having to see the youngest version of any event. Please integrate the working PRs mentioned above. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16430348#comment-16430348 ] Ben commented on KAFKA-1194: Note we also see this issue on Windows, but in slightly different circumstances. If we try to delete a topic then the broker crashes entirely with an error about not being able to delete files while they are in use. You then can't restart the broker until you delete all the log files manually first. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16379353#comment-16379353 ] David Hay commented on KAFKA-1194: -- How does one work around this issue in the meantime? We're seeing it on one of our QA servers. When the exception is thrown, the broker process abruptly terminates. Can I just delete the ".deleted" files? Is there another way to fix the problem? Kafka 0.8.2.2 CentOS > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch, windows > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16332473#comment-16332473 ] Michael van der Haven commented on KAFKA-1194: -- We use Windows deployments in closed data-room experiments and have to start-up the associated hosts a few times a day. With log-files for topics reaching their retention period, version 0.10.0.0 and 0.11.0.0 would raise an error during start-up, but they did not crash and continued to work. Using a redundant standby broker we can use a script to manually remove the old log files. Version 1.0.0.0 crashes though with the start-up and updated 'work around' scripts haven't worked-out yet. It is a bit sad that the latest version seems to have become more unstable. So I guess I'm another candidate that would really like to see this issue fixed on Windows. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[ https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16328549#comment-16328549 ] Georgij Cernysiov commented on KAFKA-1194: -- [~manme...@gmail.com] thank you. There is also [~nxmbriggs404] [pull request|https://github.com/apache/kafka/pull/3283] which fixes the issue, but there was no activity since August 2017. Recently, I merged the latest trunk into [~nxmbriggs404] branch with fixes and opened [new pull request|https://github.com/apache/kafka/pull/4431] (_i have no rights to push to his repository or to fix merge conflicts within pull requests_) which is "ready" to be merged into `trunk` (_code review required_). So far I had no issues with log retention when running kafka based on [~nxmbriggs404] fixes. > The kafka broker cannot delete the old log files after the configured time > -- > > Key: KAFKA-1194 > URL: https://issues.apache.org/jira/browse/KAFKA-1194 > Project: Kafka > Issue Type: Bug > Components: log >Affects Versions: 0.10.0.0, 0.11.0.0, 1.0.0 > Environment: window >Reporter: Tao Qin >Priority: Critical > Labels: features, patch > Attachments: KAFKA-1194.patch, Untitled.jpg, kafka-1194-v1.patch, > kafka-1194-v2.patch, screenshot-1.png > > Original Estimate: 72h > Remaining Estimate: 72h > > We tested it in windows environment, and set the log.retention.hours to 24 > hours. > # The minimum age of a log file to be eligible for deletion > log.retention.hours=24 > After several days, the kafka broker still cannot delete the old log file. > And we get the following exceptions: > [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task > 'kafka-log-retention' (kafka.utils.KafkaScheduler) > kafka.common.KafkaStorageException: Failed to change the log file suffix from > to .deleted for log segment 1516723 > at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249) > at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638) > at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418) > at > scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59) > at scala.collection.immutable.List.foreach(List.scala:76) > at kafka.log.Log.deleteOldSegments(Log.scala:418) > at > kafka.log.LogManager.kafka$log$LogManager$$cleanupExpiredSegments(LogManager.scala:284) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:316) > at > kafka.log.LogManager$$anonfun$cleanupLogs$3.apply(LogManager.scala:314) > at > scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:743) > at scala.collection.Iterator$class.foreach(Iterator.scala:772) > at > scala.collection.JavaConversions$JIteratorWrapper.foreach(JavaConversions.scala:573) > at scala.collection.IterableLike$class.foreach(IterableLike.scala:73) > at > scala.collection.JavaConversions$JListWrapper.foreach(JavaConversions.scala:615) > at > scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:742) > at kafka.log.LogManager.cleanupLogs(LogManager.scala:314) > at > kafka.log.LogManager$$anonfun$startup$1.apply$mcV$sp(LogManager.scala:143) > at kafka.utils.KafkaScheduler$$anon$1.run(KafkaScheduler.scala:100) > at > java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) > at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:304) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:178) > at > java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) > at java.lang.Thread.run(Thread.java:724) > I think this error happens because kafka tries to rename the log file when it > is still opened. So we should close the file first before rename. > The index file uses a special data structure, the MappedByteBuffer. Javadoc > describes it as: > A mapped byte buffer and the file mapping that it represents remain valid > until the buffer itself is garbage-collected. > Fortunately, I find a forceUnmap function in kafka code, and perhaps it can > be used to free the MappedByteBuffer. -- This message was sent by Atlass
[jira] [Commented] (KAFKA-1194) The kafka broker cannot delete the old log files after the configured time
[
https://issues.apache.org/jira/browse/KAFKA-1194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16079678#comment-16079678
]
M. Manna commented on KAFKA-1194:
-
I believe I have found a workaround (and possibly a solution). The Root cause
is probably on Windows FILE_SHARE_DELETE (using some internal low level API
call) is always set to false (or simply hasn't been defined). This is possibly
failing the Files.move(). Perhaps future JDKs will consider this to be
configurable somehow!.
Meanwhile, i have done the following:
1) I created a wrapper around forceUnmap as follows (in AbstractIndex.scala):
{code:java}
def forceUnmapWrapper() {
forceUnmap(mmap)
}
{code}
2) In Log.scala made the following changes for changeFileSuffixes:
{code:java}
def changeFileSuffixes(oldSuffix: String, newSuffix: String) {
def kafkaStorageException(fileType: String, e: IOException) =
new KafkaStorageException(s"Failed to change the $fileType file suffix
from $oldSuffix to $newSuffix for log segment $baseOffset", e)
logger.warn("KAFKA mod - starting log renameTo op");
try log.renameTo(new File(CoreUtils.replaceSuffix(log.file.getPath,
oldSuffix, newSuffix)))
catch {
case e: IOException => throw kafkaStorageException("log", e)
}
logger.warn("KAFKA mod - starting index renameTo op")
index.forceUnmapWrapper
try index.renameTo(new File(CoreUtils.replaceSuffix(index.file.getPath,
oldSuffix, newSuffix)))
catch {
case e: IOException => throw kafkaStorageException("index", e)
}
logger.warn("KAFKA mod - starting timeIndex renameTo op")
timeIndex.forceUnmapWrapper
try timeIndex.renameTo(new
File(CoreUtils.replaceSuffix(timeIndex.file.getPath, oldSuffix, newSuffix)))
catch {
case e: IOException => throw kafkaStorageException("timeindex", e)
}
}
{code}
Produces the following output upon startup on my 3 brokers:
{code:java}
[2017-07-09 18:42:16,451] INFO Deleting segment 0 from log z1-1. (kafka.log.Log)
[2017-07-09 18:42:16,460] INFO Deleting index
C:\tmp\kafka-logs3\z1-1\.index.deleted
(kafka.log.OffsetIndex)
[2017-07-09 18:42:16,470] INFO Deleting index
C:\tmp\kafka-logs3\z1-1\.timeindex.deleted
(kafka.log.TimeIndex)
[2017-07-09 18:42:16,484] INFO Deleting segment 30240 from log z1-1.
(kafka.log.Log)
[2017-07-09 18:42:16,501] INFO Deleting index
C:\tmp\kafka-logs3\z1-1\00030240.index.deleted
(kafka.log.OffsetIndex)
[2017-07-09 18:42:16,507] INFO Deleting index
C:\tmp\kafka-logs3\z1-1\00030240.timeindex.deleted
(kafka.log.TimeIndex)
[2017-07-09 18:42:16,517] INFO Deleting segment 47520 from log z1-1.
(kafka.log.Log)
[2017-07-09 18:42:16,520] INFO Deleting index
C:\tmp\kafka-logs3\z1-1\00047520.index.deleted
(kafka.log.OffsetIndex)
[2017-07-09 18:42:16,523] INFO Deleting index
C:\tmp\kafka-logs3\z1-1\00047520.timeindex.deleted
(kafka.log.TimeIndex)
{code}
My log.retention.minutes=10 and log.retention.check.interval.ms=30 - this
doesn't always get triggered as expected but when it does - it now cleans.
if someone is kind enough to verify this solution and propose a commit - we can
try this out for future release?
> The kafka broker cannot delete the old log files after the configured time
> --
>
> Key: KAFKA-1194
> URL: https://issues.apache.org/jira/browse/KAFKA-1194
> Project: Kafka
> Issue Type: Bug
> Components: log
>Affects Versions: 0.8.1
> Environment: window
>Reporter: Tao Qin
>Priority: Critical
> Labels: features, patch
> Attachments: KAFKA-1194.patch, kafka-1194-v1.patch,
> kafka-1194-v2.patch, screenshot-1.png, Untitled.jpg
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> We tested it in windows environment, and set the log.retention.hours to 24
> hours.
> # The minimum age of a log file to be eligible for deletion
> log.retention.hours=24
> After several days, the kafka broker still cannot delete the old log file.
> And we get the following exceptions:
> [2013-12-19 01:57:38,528] ERROR Uncaught exception in scheduled task
> 'kafka-log-retention' (kafka.utils.KafkaScheduler)
> kafka.common.KafkaStorageException: Failed to change the log file suffix from
> to .deleted for log segment 1516723
> at kafka.log.LogSegment.changeFileSuffixes(LogSegment.scala:249)
> at kafka.log.Log.kafka$log$Log$$asyncDeleteSegment(Log.scala:638)
> at kafka.log.Log.kafka$log$Log$$deleteSegment(Log.scala:629)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at kafka.log.Log$$anonfun$deleteOldSegments$1.apply(Log.scala:418)
> at
> scala.collection.LinearSeqOptimize
