----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/62388/#review185594 -----------------------------------------------------------
ambari-server/src/main/resources/common-services/YARN/2.1.0.2.0/package/scripts/service.py Lines 88 (patched) <https://reviews.apache.org/r/62388/#comment261898> Sounds like a very dirty hack. Does that reproduce in latest Ambari versions? Maybe it would be better to check status() before attempt to run start() ? - Dmitro Lisnichenko On Sept. 18, 2017, 9 p.m., Eugene Chekanskiy wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/62388/ > ----------------------------------------------------------- > > (Updated Sept. 18, 2017, 9 p.m.) > > > Review request for Ambari, Andrew Onischuk, Dmitro Lisnichenko, and Robert > Levas. > > > Bugs: AMBARI-21984 > https://issues.apache.org/jira/browse/AMBARI-21984 > > > Repository: ambari > > > Description > ------- > > Sometimes scripts starting yarn services but they are already running. It is > rarely reproduced and hard to debug. > This patch checks output of start command and ignore start failure in case > script reporting "service is running" and validates pid pile after that. > > > Diffs > ----- > > > ambari-server/src/main/resources/common-services/YARN/2.1.0.2.0/package/scripts/service.py > 7c59b60 > > ambari-server/src/main/resources/common-services/YARN/3.0.0.3.0/package/scripts/service.py > 1c612ef > ambari-server/src/test/python/stacks/2.0.6/YARN/test_historyserver.py > 6a99dfa > ambari-server/src/test/python/stacks/2.0.6/YARN/test_nodemanager.py 642043d > ambari-server/src/test/python/stacks/2.0.6/YARN/test_resourcemanager.py > 8a703df > ambari-server/src/test/python/stacks/2.1/YARN/test_apptimelineserver.py > b352920 > > > Diff: https://reviews.apache.org/r/62388/diff/1/ > > > Testing > ------- > > mvn clean test, cluster deploy > > > Thanks, > > Eugene Chekanskiy > >
