> On June 22, 2017, 4:43 p.m., Jonathan Hurley wrote: > > ambari-server/src/main/java/org/apache/ambari/server/actionmanager/ActionDBAccessorImpl.java > > Lines 837-842 (patched) > > <https://reviews.apache.org/r/60277/diff/1/?file=1757441#file1757441line837> > > > > Feels like this needs to be atomic/transactional.
org.apache.ambari.server.actionmanager.ActionDBAccessorImpl#resubmitTasks is called from org.apache.ambari.server.actionmanager.ActionManager#resubmitTasks that is called from org.apache.ambari.server.controller.internal.UpgradeResourceProvider#setUpgradeRequestStatus that has @Transactional annotation on it - Dmitro ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/60277/#review178675 ----------------------------------------------------------- On June 22, 2017, 1:20 p.m., Dmitro Lisnichenko wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/60277/ > ----------------------------------------------------------- > > (Updated June 22, 2017, 1:20 p.m.) > > > Review request for Ambari, Jonathan Hurley and Nate Cole. > > > Bugs: AMBARI-21298 > https://issues.apache.org/jira/browse/AMBARI-21298 > > > Repository: ambari > > > Description > ------- > > when upgrade is paused at a first time: > upgrade is paused > request:ABORTED > Stages (all):ABORTED > host_role_commands(all):ABORTED > > If I resume upgrade and then stop it again, often (not always) I see the > following state: > upgrade is paused, current command shows up as failed (with cancel/retry) > request:ABORTED > Stages (all):ABORTED > host_role_commands(current):HOLDING_FAILED > host_role_commands(all):ABORTED > > Resume does not transition ABORTED stages back to pending > > > Diffs > ----- > > > ambari-server/src/main/java/org/apache/ambari/server/actionmanager/ActionDBAccessorImpl.java > 94bb8926fc > > > Diff: https://reviews.apache.org/r/60277/diff/1/ > > > Testing > ------- > > tested on live cluster, seems to work well > tests passed > > > Thanks, > > Dmitro Lisnichenko > >
