[FLINK-16510] Task manager safeguard shutdown may not be reliable - ASF JIRA

Attach files

Attach Screenshot

Voters

Watch issue

Watchers

Create sub-task

Link

Clone

Update Comment Author

Replace String in Comment

Update Comment Visibility

Delete Comments

XML

Word

Printable

JSON

Details

Type: Bug
Status: Closed
Priority: Major
Resolution: Fixed
Affects Version/s: 1.10.1, 1.11.1, 1.12.0
Fix Version/s: 1.10.2, 1.11.2, 1.12.0
Component/s: Runtime / Task
Labels:
- pull-request-available

Description

The JvmShutdownSafeguard does not always succeed but can hang when multiple threads attempt to shutdown the JVM. Apparently mixing System.exit() with ShutdownHooks and forcefully terminating the JVM via Runtime.halt() does not play together well:

"Jvm Terminator" #22 daemon prio=5 os_prio=0 tid=0x00007fb8e82f2800 nid=0x5a96 runnable [0x00007fb35cffb000]
   java.lang.Thread.State: RUNNABLE
	at java.lang.Shutdown.$$YJP$$halt0(Native Method)
	at java.lang.Shutdown.halt0(Shutdown.java)
	at java.lang.Shutdown.halt(Shutdown.java:139)
	- locked <0x000000047ed67638> (a java.lang.Shutdown$Lock)
	at java.lang.Runtime.halt(Runtime.java:276)
	at org.apache.flink.runtime.util.JvmShutdownSafeguard$DelayedTerminator.run(JvmShutdownSafeguard.java:86)
	at java.lang.Thread.run(Thread.java:748)

   Locked ownable synchronizers:
	- None

"FlinkCompletableFutureDelayScheduler-thread-1" #18154 daemon prio=5 os_prio=0 tid=0x00007fb708a7d000 nid=0x5a8a waiting for monitor entry [0x00007fb289d49000]
   java.lang.Thread.State: BLOCKED (on object monitor)
	at java.lang.Shutdown.halt(Shutdown.java:139)
	- waiting to lock <0x000000047ed67638> (a java.lang.Shutdown$Lock)
	at java.lang.Shutdown.exit(Shutdown.java:213)
	- locked <0x000000047edb7348> (a java.lang.Class for java.lang.Shutdown)
	at java.lang.Runtime.exit(Runtime.java:110)
	at java.lang.System.exit(System.java:973)
	at org.apache.flink.runtime.taskexecutor.TaskManagerRunner.terminateJVM(TaskManagerRunner.java:266)
	at org.apache.flink.runtime.taskexecutor.TaskManagerRunner.lambda$onFatalError$1(TaskManagerRunner.java:260)
	at org.apache.flink.runtime.taskexecutor.TaskManagerRunner$$Lambda$27464/1464672548.accept(Unknown Source)
	at java.util.concurrent.CompletableFuture.uniWhenComplete(CompletableFuture.java:774)
	at java.util.concurrent.CompletableFuture$UniWhenComplete.tryFire(CompletableFuture.java:750)
	at java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:488)
	at java.util.concurrent.CompletableFuture.completeExceptionally(CompletableFuture.java:1990)
	at org.apache.flink.runtime.concurrent.FutureUtils$Timeout.run(FutureUtils.java:943)
	at org.apache.flink.runtime.concurrent.DirectExecutorService.execute(DirectExecutorService.java:211)
	at org.apache.flink.runtime.concurrent.FutureUtils.lambda$orTimeout$11(FutureUtils.java:361)
	at org.apache.flink.runtime.concurrent.FutureUtils$$Lambda$27435/159015392.run(Unknown Source)
	at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
	at java.util.concurrent.FutureTask.run(FutureTask.java:266)
	at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
	at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
	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)

   Locked ownable synchronizers:
	- <0x00000006d5e56bd0> (a java.util.concurrent.ThreadPoolExecutor$Worker)

Note that under this condition the JVM should terminate but it still hangs. Sometimes it quits after several minutes.