[
https://issues.apache.org/jira/browse/HIVE-20602?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16775656#comment-16775656
]
t oo commented on HIVE-20602:
-----------------------------
workaround is set _hive_.metastore.event.db._notification_._api_.auth to false
> hive3 crashes after 1min
> ------------------------
>
> Key: HIVE-20602
> URL: https://issues.apache.org/jira/browse/HIVE-20602
> Project: Hive
> Issue Type: Bug
> Components: HiveServer2, Metastore, Standalone Metastore
> Affects Versions: 3.0.0
> Reporter: t oo
> Priority: Blocker
>
> Running hiveserver2 process (v3.0.0 of hive) on ec2 (not emr), the process
> starts up and for the first 1min everything is ok (I can make beeline
> connection, create/repair/select external hive tables) but then the
> hiveserver2 process crashes. If I restart the process and even do nothing the
> hiveserver2 process crashes after 1min. When checking the logs I see messages
> like 'number of connections to metastore: 1','number of connections to
> metastore: 2','number of connections to metastore: 3' then 'could not bind to
> port 10000 port already in use' then end of the logs.
> I made some experiments on few different ec2s (if i use hive v2.3.2 the
> hiveserver2 process never crashes), but if i use hive v3.0.0 it consistently
> crashes after a min.
> Metastore db is mysql rds, hive metastore process never crashed. I can see
> the external hive table ddls are persisted in the mysql (ie DBS, TBLS tables).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
