[
https://issues.apache.org/jira/browse/IGNITE-169?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15139627#comment-15139627
]
Andrey Gura edited comment on IGNITE-169 at 2/9/16 9:57 PM:
------------------------------------------------------------
I see two options for fix this:
# We can fix test. Communication SPI should be replaced. It should catch job
requests (e.g. first one), get task session, invoke {{saveCheckpoint}} method
and count latch down.
# From my point of view isn't good idea try to process delayed results from
caller thread (at least in case of async compute facade). So I suggest remove
this behavior (remove {{GridTaskWorker.processDelayedResponses}} method).
Thoughts?
was (Author: agura):
I see two options for fix this:
# We can fix test. Communication SPI should be replaced. It should cacth job
requests (e.g. first one), get task session, invoke saveCheckpoint and count
latch down.
# From my point of view isn't good idea try to process delayed results from
caller thread (especially in case of async compute facade). So I suggest remove
this bahavior (remove {{processDelayedResponses}} method).
Thoughts?
> [Failed test] GridSessionCheckpointSelfTest fails
> -------------------------------------------------
>
> Key: IGNITE-169
> URL: https://issues.apache.org/jira/browse/IGNITE-169
> Project: Ignite
> Issue Type: Test
> Components: general
> Reporter: Irina Vasilinets
> Fix For: 1.6
>
>
> {{GridSessionCheckpointSelfTest.testSharedFsCheckpoint fails}} from time to
> time.
> The problem is that test-runner thread invokes {{ignite.compute()}} in async
> mode and when all jobs are mapped it invokes {{processDelayedResponses}}
> method. At this moment is possible that all responses are received but sys
> pool threads don't invoke {{reduce}} method. test-runner thread polls delayed
> responses from queue and invoke {{reduce}}. So test-runner thread "steals"
> reduce job that is waitiong on latch that should be counted down by the same
> thread.
> Just run test N times in a loop in order to reproduce it.
> {noformat}
> [19:04:17,405][ERROR][main][root] Test failed.
> class org.apache.ignite.IgniteException: Failed to save checkpoint (session
> closed): GridTaskSessionImpl [taskName=GridCheckpointTestTask,
> dep=GridDeployment [ts=1454958227376, depMode=SHARED,
> clsLdr=IsolatedClassLoader{roleName='test'},
> clsLdrId=16c7442c251-88ba4909-a876-4106-be63-3fec03d9dcff, userVer=0,
> loc=true,
> sampleClsName=org.apache.ignite.session.GridSessionCheckpointAbstractSelfTest$GridCheckpointTestTask,
> pendingUndeploy=false, undeployed=false, usage=0],
> taskClsName=org.apache.ignite.session.GridSessionCheckpointAbstractSelfTest$GridCheckpointTestTask,
> sesId=36c7442c251-88ba4909-a876-4106-be63-3fec03d9dcff,
> startTime=1454958227376, endTime=9223372036854775807,
> taskNodeId=88ba4909-a876-4106-be63-3fec03d9dcff,
> clsLdr=IsolatedClassLoader{roleName='test'}, closed=true, cpSpi=null,
> failSpi=null, loadSpi=null, usage=0, fullSup=true,
> subjId=88ba4909-a876-4106-be63-3fec03d9dcff, mapFut=IgniteFuture
> [orig=GridFutureAdapter [resFlag=2, res=null, startTime=1454958227376,
> endTime=1454958227386, ignoreInterrupts=false, state=DONE]]]
> at
> org.apache.ignite.internal.GridTaskSessionImpl.saveCheckpoint0(GridTaskSessionImpl.java:697)
> at
> org.apache.ignite.internal.GridTaskSessionImpl.saveCheckpoint(GridTaskSessionImpl.java:677)
> at
> org.apache.ignite.internal.GridTaskSessionImpl.saveCheckpoint(GridTaskSessionImpl.java:671)
> at
> org.apache.ignite.internal.GridTaskSessionImpl.saveCheckpoint(GridTaskSessionImpl.java:666)
> at
> org.apache.ignite.session.GridSessionCheckpointAbstractSelfTest.checkCheckpoints(GridSessionCheckpointAbstractSelfTest.java:145)
> at
> org.apache.ignite.session.GridSessionCheckpointSelfTest.testSharedFsCheckpoint(GridSessionCheckpointSelfTest.java:48)
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> at java.lang.reflect.Method.invoke(Method.java:606)
> at junit.framework.TestCase.runTest(TestCase.java:176)
> at
> org.apache.ignite.testframework.junits.GridAbstractTest.runTestInternal(GridAbstractTest.java:1723)
> at
> org.apache.ignite.testframework.junits.GridAbstractTest.access$000(GridAbstractTest.java:118)
> at
> org.apache.ignite.testframework.junits.GridAbstractTest$4.run(GridAbstractTest.java:1661)
> at java.lang.Thread.run(Thread.java:745)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
