[KAFKA-6322] Error deleting log for topic, all log dirs failed.

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.0.0
• Fix Version/s: 1.1.0
• Environment: RancherOS with NFS mounted for log directory
• Assignee: Ismael Juma
• Reporter: dongyan li
• Created: 06/Dec/17 22:46
• Updated: 12/Dec/17 08:04
• Resolved: 12/Dec/17 08:04

Description

Hello,

I encountered a error when I try to delete a topic with kafka version 1.0.0, the error is not present on version 0.10.2.1 which is the version I upgraded from.

I suspect that some other thread is still using that file while the Kafka is trying to delete that.

12/6/2017 3:37:32 PM[2017-12-06 21:37:32,846] ERROR Exception while deleting Log(/opt/kafka/logs/topicname-0.112ff11872e4411ca7470ba7e3026ab0-delete) in dir /opt/kafka/logs. (kafka.log.LogManager)
12/6/2017 3:37:32 PMorg.apache.kafka.common.errors.KafkaStorageException: Error while deleting log for topicname-0 in dir /opt/kafka/logs
12/6/2017 3:37:32 PMCaused by: java.nio.file.FileSystemException: /opt/kafka/logs/topicname-0.112ff11872e4411ca7470ba7e3026ab0-delete/.nfs0000000000e609f2000000ce: Device or resource busy
12/6/2017 3:37:32 PM	at sun.nio.fs.UnixException.translateToIOException(UnixException.java:91)
12/6/2017 3:37:32 PM	at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:102)
12/6/2017 3:37:32 PM	at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:107)
12/6/2017 3:37:32 PM	at sun.nio.fs.UnixFileSystemProvider.implDelete(UnixFileSystemProvider.java:244)
12/6/2017 3:37:32 PM	at sun.nio.fs.AbstractFileSystemProvider.delete(AbstractFileSystemProvider.java:103)
12/6/2017 3:37:32 PM	at java.nio.file.Files.delete(Files.java:1126)
12/6/2017 3:37:32 PM	at org.apache.kafka.common.utils.Utils$1.visitFile(Utils.java:630)
12/6/2017 3:37:32 PM	at org.apache.kafka.common.utils.Utils$1.visitFile(Utils.java:619)
12/6/2017 3:37:32 PM	at java.nio.file.Files.walkFileTree(Files.java:2670)
12/6/2017 3:37:32 PM	at java.nio.file.Files.walkFileTree(Files.java:2742)
12/6/2017 3:37:32 PM	at org.apache.kafka.common.utils.Utils.delete(Utils.java:619)
12/6/2017 3:37:32 PM	at kafka.log.Log.$anonfun$delete$2(Log.scala:1432)
12/6/2017 3:37:32 PM	at scala.runtime.java8.JFunction0$mcV$sp.apply(JFunction0$mcV$sp.java:12)
12/6/2017 3:37:32 PM	at kafka.log.Log.maybeHandleIOException(Log.scala:1669)
12/6/2017 3:37:32 PM	at kafka.log.Log.delete(Log.scala:1427)
12/6/2017 3:37:32 PM	at kafka.log.LogManager.deleteLogs(LogManager.scala:626)
12/6/2017 3:37:32 PM	at kafka.log.LogManager.$anonfun$startup$7(LogManager.scala:362)
12/6/2017 3:37:32 PM	at kafka.utils.KafkaScheduler.$anonfun$schedule$2(KafkaScheduler.scala:110)
12/6/2017 3:37:32 PM	at kafka.utils.CoreUtils$$anon$1.run(CoreUtils.scala:61)
12/6/2017 3:37:32 PM	at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
12/6/2017 3:37:32 PM	at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308)
12/6/2017 3:37:32 PM	at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180)
12/6/2017 3:37:32 PM	at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294)
12/6/2017 3:37:32 PM	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
12/6/2017 3:37:32 PM	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
12/6/2017 3:37:32 PM	at java.lang.Thread.run(Thread.java:748)

Then

12/6/2017 3:37:32 PM[2017-12-06 21:37:32,882] INFO Stopping serving logs in dir /opt/kafka/logs (kafka.log.LogManager)
12/6/2017 3:37:32 PM[2017-12-06 21:37:32,883] FATAL Shutdown broker because all log dirs in /opt/kafka/logs have failed (kafka.log.LogManager)