[jira] [Comment Edited] (SPARK-30460) Spark checkpoint failing after some run with S3 path
[
https://issues.apache.org/jira/browse/SPARK-30460?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17012784#comment-17012784
]
Sachin Pasalkar edited comment on SPARK-30460 at 1/10/20 12:19 PM:
---
[~gsomogyi] Yes may be or may be not.
I was able to run this on my production for 4-6 hours without any other issues
for 4-5 times. It always failed with this issue. If this fix the some part of
problem we should fix it.
I understand spark 3.0 has new committer but as you said it is not deeply
tested. Soon I am going to run my Production with this fix in place, I will
update ticket around next EOW. If I was able to run system smoothly or not
was (Author: sachin):
Yes may be or may be not.
I was able to run this on my production for 4-6 hours without any other issues
for 4-5 times. It always failed with this issue. If this fix the some part of
problem we should fix it.
I understand spark 3.0 has new committer but as you said it is not deeply
tested. Soon I am going to run my Production with this fix in place, I will
update ticket around next EOW. If I was able to run system smoothly or not
> Spark checkpoint failing after some run with S3 path
> -
>
> Key: SPARK-30460
> URL: https://issues.apache.org/jira/browse/SPARK-30460
> Project: Spark
> Issue Type: Bug
> Components: DStreams
>Affects Versions: 2.4.4
>Reporter: Sachin Pasalkar
>Priority: Major
>
> We are using EMR with the SQS as source of stream. However it is failing,
> after 4-6 hours of run, with below exception. Application shows its running
> but stops the processing the messages
> {code:java}
> 2020-01-06 13:04:10,548 WARN [BatchedWriteAheadLog Writer]
> org.apache.spark.streaming.util.BatchedWriteAheadLog:BatchedWriteAheadLog
> Writer failed to write ArrayBuffer(Record(java.nio.HeapByteBuffer[pos=0
> lim=1226 cap=1226],1578315850302,Future()))
> java.lang.UnsupportedOperationException
> at
> com.amazon.ws.emr.hadoop.fs.s3n2.S3NativeFileSystem2.append(S3NativeFileSystem2.java:150)
> at org.apache.hadoop.fs.FileSystem.append(FileSystem.java:1181)
> at
> com.amazon.ws.emr.hadoop.fs.EmrFileSystem.append(EmrFileSystem.java:295)
> at
> org.apache.spark.streaming.util.HdfsUtils$.getOutputStream(HdfsUtils.scala:35)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLogWriter.stream$lzycompute(FileBasedWriteAheadLogWriter.scala:32)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLogWriter.stream(FileBasedWriteAheadLogWriter.scala:32)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLogWriter.(FileBasedWriteAheadLogWriter.scala:35)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLog.getLogWriter(FileBasedWriteAheadLog.scala:229)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLog.write(FileBasedWriteAheadLog.scala:94)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLog.write(FileBasedWriteAheadLog.scala:50)
> at
> org.apache.spark.streaming.util.BatchedWriteAheadLog.org$apache$spark$streaming$util$BatchedWriteAheadLog$$flushRecords(BatchedWriteAheadLog.scala:175)
> at
> org.apache.spark.streaming.util.BatchedWriteAheadLog$$anon$1.run(BatchedWriteAheadLog.scala:142)
> at java.lang.Thread.run(Thread.java:748)
> 2020-01-06 13:04:10,554 WARN [wal-batching-thread-pool-0]
> org.apache.spark.streaming.scheduler.ReceivedBlockTracker:Exception thrown
> while writing record:
> BlockAdditionEvent(ReceivedBlockInfo(0,Some(3),None,WriteAheadLogBasedStoreResult(input-0-1578315849800,Some(3),FileBasedWriteAheadLogSegment(s3://mss-prod-us-east-1-ueba-bucket/streaming/checkpoint/receivedData/0/log-1578315850001-1578315910001,0,5175
> to the WriteAheadLog.
> org.apache.spark.SparkException: Exception thrown in awaitResult:
> at org.apache.spark.util.ThreadUtils$.awaitResult(ThreadUtils.scala:226)
> at
> org.apache.spark.streaming.util.BatchedWriteAheadLog.write(BatchedWriteAheadLog.scala:84)
> at
> org.apache.spark.streaming.scheduler.ReceivedBlockTracker.writeToLog(ReceivedBlockTracker.scala:242)
> at
> org.apache.spark.streaming.scheduler.ReceivedBlockTracker.addBlock(ReceivedBlockTracker.scala:89)
> at
> org.apache.spark.streaming.scheduler.ReceiverTracker.org$apache$spark$streaming$scheduler$ReceiverTracker$$addBlock(ReceiverTracker.scala:347)
> at
> org.apache.spark.streaming.scheduler.ReceiverTracker$ReceiverTrackerEndpoint$$anonfun$receiveAndReply$1$$anon$1$$anonfun$run$1.apply$mcV$sp(ReceiverTracker.scala:522)
> at org.apache.spark.util.Utils$.tryLogNonFatalError(Utils.scala:1340)
> at
>
[jira] [Comment Edited] (SPARK-30460) Spark checkpoint failing after some run with S3 path
[
https://issues.apache.org/jira/browse/SPARK-30460?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17012756#comment-17012756
]
Gabor Somogyi edited comment on SPARK-30460 at 1/10/20 11:40 AM:
-
[~Sachin] even if somebody hunt down this specific issue S3 checkpoint makes
streaming jobs dead many other ways.
was (Author: gsomogyi):
[~Sachin] even if somebody hunt down this specific issue S3 checkpoint makes
streaming jobs dead many other different ways.
> Spark checkpoint failing after some run with S3 path
> -
>
> Key: SPARK-30460
> URL: https://issues.apache.org/jira/browse/SPARK-30460
> Project: Spark
> Issue Type: Bug
> Components: DStreams
>Affects Versions: 2.4.4
>Reporter: Sachin Pasalkar
>Priority: Major
>
> We are using EMR with the SQS as source of stream. However it is failing,
> after 4-6 hours of run, with below exception. Application shows its running
> but stops the processing the messages
> {code:java}
> 2020-01-06 13:04:10,548 WARN [BatchedWriteAheadLog Writer]
> org.apache.spark.streaming.util.BatchedWriteAheadLog:BatchedWriteAheadLog
> Writer failed to write ArrayBuffer(Record(java.nio.HeapByteBuffer[pos=0
> lim=1226 cap=1226],1578315850302,Future()))
> java.lang.UnsupportedOperationException
> at
> com.amazon.ws.emr.hadoop.fs.s3n2.S3NativeFileSystem2.append(S3NativeFileSystem2.java:150)
> at org.apache.hadoop.fs.FileSystem.append(FileSystem.java:1181)
> at
> com.amazon.ws.emr.hadoop.fs.EmrFileSystem.append(EmrFileSystem.java:295)
> at
> org.apache.spark.streaming.util.HdfsUtils$.getOutputStream(HdfsUtils.scala:35)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLogWriter.stream$lzycompute(FileBasedWriteAheadLogWriter.scala:32)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLogWriter.stream(FileBasedWriteAheadLogWriter.scala:32)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLogWriter.(FileBasedWriteAheadLogWriter.scala:35)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLog.getLogWriter(FileBasedWriteAheadLog.scala:229)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLog.write(FileBasedWriteAheadLog.scala:94)
> at
> org.apache.spark.streaming.util.FileBasedWriteAheadLog.write(FileBasedWriteAheadLog.scala:50)
> at
> org.apache.spark.streaming.util.BatchedWriteAheadLog.org$apache$spark$streaming$util$BatchedWriteAheadLog$$flushRecords(BatchedWriteAheadLog.scala:175)
> at
> org.apache.spark.streaming.util.BatchedWriteAheadLog$$anon$1.run(BatchedWriteAheadLog.scala:142)
> at java.lang.Thread.run(Thread.java:748)
> 2020-01-06 13:04:10,554 WARN [wal-batching-thread-pool-0]
> org.apache.spark.streaming.scheduler.ReceivedBlockTracker:Exception thrown
> while writing record:
> BlockAdditionEvent(ReceivedBlockInfo(0,Some(3),None,WriteAheadLogBasedStoreResult(input-0-1578315849800,Some(3),FileBasedWriteAheadLogSegment(s3://mss-prod-us-east-1-ueba-bucket/streaming/checkpoint/receivedData/0/log-1578315850001-1578315910001,0,5175
> to the WriteAheadLog.
> org.apache.spark.SparkException: Exception thrown in awaitResult:
> at org.apache.spark.util.ThreadUtils$.awaitResult(ThreadUtils.scala:226)
> at
> org.apache.spark.streaming.util.BatchedWriteAheadLog.write(BatchedWriteAheadLog.scala:84)
> at
> org.apache.spark.streaming.scheduler.ReceivedBlockTracker.writeToLog(ReceivedBlockTracker.scala:242)
> at
> org.apache.spark.streaming.scheduler.ReceivedBlockTracker.addBlock(ReceivedBlockTracker.scala:89)
> at
> org.apache.spark.streaming.scheduler.ReceiverTracker.org$apache$spark$streaming$scheduler$ReceiverTracker$$addBlock(ReceiverTracker.scala:347)
> at
> org.apache.spark.streaming.scheduler.ReceiverTracker$ReceiverTrackerEndpoint$$anonfun$receiveAndReply$1$$anon$1$$anonfun$run$1.apply$mcV$sp(ReceiverTracker.scala:522)
> at org.apache.spark.util.Utils$.tryLogNonFatalError(Utils.scala:1340)
> at
> org.apache.spark.streaming.scheduler.ReceiverTracker$ReceiverTrackerEndpoint$$anonfun$receiveAndReply$1$$anon$1.run(ReceiverTracker.scala:520)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> at java.lang.Thread.run(Thread.java:748)
> Caused by: java.lang.UnsupportedOperationException
> at
> com.amazon.ws.emr.hadoop.fs.s3n2.S3NativeFileSystem2.append(S3NativeFileSystem2.java:150)
> at org.apache.hadoop.fs.FileSystem.append(FileSystem.java:1181)
> at
> com.amazon.ws.emr.hadoop.fs.EmrFileSystem.append(EmrFileSystem.java:295)
> at
>
