[jira] [Commented] (AMBARI-17799) AMS: Deleting ZK data in embedded mode results in external phoenix clients unable to connect on AMS restart

["Di Li (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/AMBARI-17799?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15648135#comment-15648135 ] 

Di Li commented on AMBARI-17799:
--------------------------------

Hello Sid, 

Sorry to bother you, I hit the AMS collector cannot restart (due to corrupted zookeeper datafile) issue on a single node Ambari 2.5/HDP 2.6 cluster. Since the removing zookeeper data dir logic was removed via this JIRA, I wonder if AMS HBase has been updated to include the fix. I am still getting the HBase master not active after 30sec error message.

Thanks.

Di

> AMS: Deleting ZK data in embedded mode results in external phoenix clients unable to connect on AMS restart
> -----------------------------------------------------------------------------------------------------------
>
>                 Key: AMBARI-17799
>                 URL: https://issues.apache.org/jira/browse/AMBARI-17799
>             Project: Ambari
>          Issue Type: Bug
>          Components: ambari-metrics
>    Affects Versions: 2.0.0
>            Reporter: Siddharth Wagle
>            Assignee: Siddharth Wagle
>            Priority: Critical
>             Fix For: 2.4.0
>
>         Attachments: AMBARI-17799.patch
>
>
> AMS stack script delete ZK data because when standalone HBase restarts, the ZK data quite frequently gets corrupt and the restart is no longer works. 
> This is because all daemon run inside the same JVM. We had to do this as a workaround after which restart of AMS in embedded mode became reliable
> With HBASE-16244, the standalone mode restart will succeed and we do not need to cleanup zookeeper data



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)