Ian Downes created MESOS-1189:
---------------------------------
Summary: Slave::shutdownExecutor normally doesn't destroy
containers
Key: MESOS-1189
URL: https://issues.apache.org/jira/browse/MESOS-1189
Project: Mesos
Issue Type: Bug
Components: isolation, test
Affects Versions: 0.18.0
Reporter: Ian Downes
Priority: Minor
Fix For: 0.19.0
Slave::shutdownExecutor requests the executor to shut down and will only
explicitly destroy the container on timeout. Normally this is okay because
MesosContainerizer will automatically destroy the container when it notices the
executor pid has exited but in some tests the containerizer is destroyed before
this happens, leading to defunct executor processes which will be picked up by
the test environment checks if the reaper hasn't yet waited on them (which is
quite likely given the reap period and test execution speed).
{noformat}
[ OK ] SlaveRecoveryTest/0.RemoveNonCheckpointingFramework (1536 ms)
[----------] 1 test from SlaveRecoveryTest/0 (1536 ms total)
[----------] Global test environment tear-down
../../src/tests/environment.cpp:244: Failure
Failed
Tests completed with child processes remaining:
-+- 6045 /home/idownes/workspace/mesos/build/src/.libs/lt-mesos-tests
--gtest_filter=SlaveRecoveryTest/0.RemoveNonCheckpointingFramework
|--- 6086 ()
\--- 6099 ()
[==========] 1 test from 1 test case ran. (1549 ms total)
[ PASSED ] 1 test.
[ FAILED ] 0 tests, listed below:
{noformat}
--
This message was sent by Atlassian JIRA
(v6.2#6252)
