[ https://issues.apache.org/jira/browse/HIVE-19597?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16482784#comment-16482784 ]
Sergey Shelukhin commented on HIVE-19597: ----------------------------------------- Update: (I am not looking at this for now but plan to look this week...) it happens when we get a session on a separate thread and then join() it. I added logging and reproed locally and it looked like the spawned thread got the session but didn't exit, which is not supposed to be possible. I will look more later. > TestWorkloadManager sometimes hangs > ----------------------------------- > > Key: HIVE-19597 > URL: https://issues.apache.org/jira/browse/HIVE-19597 > Project: Hive > Issue Type: Bug > Reporter: Sergey Shelukhin > Assignee: Sergey Shelukhin > Priority: Major > > Seems like the tests randomly get stuck after the lines like > {noformat} > 2018-05-17T01:54:27,111 INFO [Workload management master] > tez.WorkloadManager: Processing current events > 2018-05-17T01:54:27,603 INFO [TriggerValidator] > tez.PerPoolTriggerValidatorRunnable: Creating trigger validator for pool: llap > 2018-05-17T01:54:37,090 DEBUG [Thread-28] conf.HiveConf: Found metastore URI > of null > {noformat} > Then they get killed by timeout. Happened in the same manner, to random tests > in a few separate runs. -- This message was sent by Atlassian JIRA (v7.6.3#76005)