[
https://issues.apache.org/jira/browse/IMPALA-11291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17538154#comment-17538154
]
ASF subversion and git services commented on IMPALA-11291:
----------------------------------------------------------
Commit 9f2447b1b382538404057fa9572a693f2215378c in impala's branch
refs/heads/master from stiga-huang
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=9f2447b1b ]
IMPALA-11291: deflake minidump-test by not starting JVM
minidump-test is unified into unified-be-test which will always init the
JVM and start the JVM pause monitor. It will print the following log to
stdout:
22/05/16 11:35:46 INFO util.JvmPauseMonitor: Starting JVM pause monitor
It's printed concurrently with the minidump output, which makes the test
flaky since it's verifying the output lines.
This patch moves minidump-test out of unified-be-test and explicitly
invokes InitCommonRuntime with init_jvm=false. So it won't be disrupted
by the above log.
Tests
- The issue occurs when I run minidump-test several times repeatedly.
After the fix, I can run minidump-test more than 2000 times without
errors.
Change-Id: I89f81d408a2e905d5dfdd7f87177ebe2079d4d27
Reviewed-on: http://gerrit.cloudera.org:8080/18529
Reviewed-by: Impala Public Jenkins <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
> minidump-test is flaky
> ----------------------
>
> Key: IMPALA-11291
> URL: https://issues.apache.org/jira/browse/IMPALA-11291
> Project: IMPALA
> Issue Type: Bug
> Affects Versions: Impala 4.2.0
> Reporter: Quanlong Huang
> Assignee: Michael Smith
> Priority: Critical
>
> Minidump.DumpCallback and Minidump.DumpCallbackWithThread added by
> IMPALA-11275 are flaky.
> Minidump.DumpCallback is found to fail in an unrelated build:
> https://jenkins.impala.io/job/ubuntu-16.04-from-scratch/16449/testReport/junit/(root)/Minidump/DumpCallback/
> {noformat}
> /home/ubuntu/Impala/be/src/util/minidump-test.cc:47
> Value of: lines.size()
> Actual: 4
> Expected: 3
> {noformat}
> While looping the test locally, I can also find failures of
> DumpCallbackWithThread:
> {code}
> seed = 1652667209
> Note: Google Test filter = Minidump.*
> [==========] Running 2 tests from 1 test case.
> [----------] Global test environment set-up.
> [----------] 2 tests from Minidump
> [ RUN ] Minidump.DumpCallback
> [ OK ] Minidump.DumpCallback (1 ms)
> [ RUN ] Minidump.DumpCallbackWithThread
> /home/quanlong/workspace/Impala/be/src/util/minidump-test.cc:76: Failure
> Value of: lines.size()
> Actual: 4
> Expected: 3
> /home/quanlong/workspace/Impala/be/src/util/minidump-test.cc:77: Failure
> Value of: boost::regex_match(lines[0], minidump_in_thread)
> Actual: false
> Expected: true
> /home/quanlong/workspace/Impala/be/src/util/minidump-test.cc:78: Failure
> Value of: boost::regex_match(lines[1], wrote_minidump)
> Actual: false
> Expected: true
> /home/quanlong/workspace/Impala/be/src/util/minidump-test.cc:79: Failure
> Value of: lines[2]
> Actual: "Wrote minidump to
> /tmp/arbitrary/path/b5cfb1b6-14cc-4a43-676f738b-52e07232.dmp"
> Expected: ""
> [ FAILED ] Minidump.DumpCallbackWithThread (0 ms)
> [----------] 2 tests from Minidump (1 ms total)
> [----------] Global test environment tear-down
> [==========] 2 tests from 1 test case ran. (1 ms total)
> [ PASSED ] 1 test.
> [ FAILED ] 1 test, listed below:
> [ FAILED ] Minidump.DumpCallbackWithThread
> {code}
> CC [~MikaelSmith]
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
