[
https://issues.apache.org/jira/browse/HIVE-12797?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vikram Dixit K updated HIVE-12797:
----------------------------------
Attachment: HIVE-12797.2.patch
[~sershe] It looks like we only need the synchronization for the iterations
according to the javadoc. I added the synchronized block in the TezJobMonitor
and while at it found another synchronizedList being used in a similar fashion
inside that class and fixed that as well. I can remove this synchronizedList
business fully as it actually does the same thing as using a synchronized block
internally in its add/remove methods. I am leaning towards the solution in this
patch for now. Let me know if you feel strongly towards removing this
synchronized list and synchronize manually.
> Synchronization issues with tez/llap session pool in hs2
> --------------------------------------------------------
>
> Key: HIVE-12797
> URL: https://issues.apache.org/jira/browse/HIVE-12797
> Project: Hive
> Issue Type: Bug
> Components: HiveServer2
> Affects Versions: 2.0.0
> Reporter: Vikram Dixit K
> Assignee: Vikram Dixit K
> Attachments: HIVE-12797.1.patch, HIVE-12797.2.patch
>
>
> The changes introduced as part of HIVE-12674 causes issues while shutting
> down hs2 when session pools are used.
> {code}
> java.util.ConcurrentModificationException
> at
> java.util.LinkedList$ListItr.checkForComodification(LinkedList.java:966)
> ~[?:1.8.0_45]
> at java.util.LinkedList$ListItr.remove(LinkedList.java:921)
> ~[?:1.8.0_45]
> at
> org.apache.hadoop.hive.ql.exec.tez.TezSessionPoolManager.stop(TezSessionPoolManager.java:288)
> ~[hive-exec-2.0.0.2.3.5.0-79.jar:2.0.0.2.3.5.0-79]
> at
> org.apache.hive.service.server.HiveServer2.stop(HiveServer2.java:479)
> [hive-jdbc-2.0.0.2.3.5.0-79-standalone.jar:2.0.0.2.3.5.0-79]
> at
> org.apache.hive.service.server.HiveServer2$2.run(HiveServer2.java:183)
> [hive-jdbc-2.0.0.2.3.5.0-79-standalone.jar:2.0.0.2.3.5.0-79]
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
