[
https://issues.apache.org/jira/browse/KAFKA-2024?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14371177#comment-14371177
]
Rajini Sivaram commented on KAFKA-2024:
---------------------------------------
The attached patch stops grouping of segments in LogCleaner if the difference
between the first offset in the group and last offset in the next segment is
greater than Int.MaxValue. Unit test for border cases included in the patch.
Also ran a manual test to recreate the reported failure in Kafka restart after
a large range of offsets were grouped together. Reran the test with the fix to
check that it fixes the issue
> Cleaner can generate unindexable log segments
> ---------------------------------------------
>
> Key: KAFKA-2024
> URL: https://issues.apache.org/jira/browse/KAFKA-2024
> Project: Kafka
> Issue Type: Bug
> Affects Versions: 0.8.2.0
> Reporter: Gian Merlino
> Attachments: KAFKA-2024.patch
>
>
> It's possible for log cleaning to generate segments that have a gap of more
> than Int.MaxValue between their base offset and their last offset. It's not
> possible to index those segments since there's only 4 bytes available to
> store that difference. The broker will end up writing overflowed ints into
> the index, and doesn't detect that there is a problem until restarted, at
> which point you get one of these:
> 2015-03-16 20:35:49,632 FATAL [main] kafka.server.KafkaServerStartable -
> Fatal error during KafkaServerStartable startup. Prepare to shutdown
> java.lang.IllegalArgumentException: requirement failed: Corrupt index found,
> index file (/mnt/persistent/kafka-logs/topic/00000000000000000000.index) has
> non-zero size but the last offset is -1634293959 and the base offset is 0
> at scala.Predef$.require(Predef.scala:233)
> at kafka.log.OffsetIndex.sanityCheck(OffsetIndex.scala:352)
> at kafka.log.Log$$anonfun$loadSegments$5.apply(Log.scala:204)
> at kafka.log.Log$$anonfun$loadSegments$5.apply(Log.scala:203)
> at scala.collection.Iterator$class.foreach(Iterator.scala:727)
> at scala.collection.AbstractIterator.foreach(Iterator.scala:1157)
> at scala.collection.IterableLike$class.foreach(IterableLike.scala:72)
> at scala.collection.AbstractIterable.foreach(Iterable.scala:54)
> at kafka.log.Log.loadSegments(Log.scala:203)
> at kafka.log.Log.<init>(Log.scala:67)
> at
> kafka.log.LogManager$$anonfun$loadLogs$2$$anonfun$3$$anonfun$apply$7$$anonfun$apply$1.apply$mcV$sp(LogManager.scala:142)
> at kafka.utils.Utils$$anon$1.run(Utils.scala:54)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
> at java.util.concurrent.FutureTask.run(FutureTask.java:262)
> 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:745)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
