[ 
https://issues.apache.org/jira/browse/HBASE-7865?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13580194#comment-13580194
 ] 

Ted Yu commented on HBASE-7865:
-------------------------------

Looking at hbase-root-master-hbasetest.out.4, connection loss to zookeeper 
happened around 06:10 :
{code}
13/02/17 06:10:46 INFO client.HConnectionManager$HConnectionImplementation: 
This client just lost it's session with ZooKeeper, will automatically reconnect 
when needed.
13/02/17 06:10:46 INFO client.HConnectionManager$HConnectionImplementation: ZK 
session expired. This disconnect could have been caused by a network partition 
or a long-running GC pause, either way it's recommended that you verify your 
environment.
Exception in thread "main" java.lang.NullPointerException
  at 
org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation.checkIfBaseNodeAvailable(HConnectionManager.java:747)
  at 
org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation.getMaster(HConnectionManager.java:685)
{code}
I went through jstack and didn't find much clue.
                
> HBase regionserver never stops when running `bin/stop-hbase.sh` on master
> -------------------------------------------------------------------------
>
>                 Key: HBASE-7865
>                 URL: https://issues.apache.org/jira/browse/HBASE-7865
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 0.94.5
>            Reporter: Jean-Marc Spaggiari
>         Attachments: hbase-root-master-hbasetest.log, 
> hbase-root-master-hbasetest.out.4, jstack_hbasetest.txt, jstack_node1.txt, 
> jstack_node3.txt, jstack_node7.txt
>
>
> I faced 3 regions (out of 8) never stopping today. This is pretty bad because 
> the script is supposed to wait until all the RS stopped to re-start 
> everything, therefor, servers are never going back online.
> HBASE-7838 will help with that and will kill the RSs. But that will not 
> really solve the root cause.
> Attached are the jstack for the 3 servers.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Reply via email to