[jira] [Commented] (HBASE-13553) ChaosMonkey should inject lease recovery failures

["Sean Busbey (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HBASE-13553?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14511181#comment-14511181 ] 

Sean Busbey commented on HBASE-13553:
-------------------------------------

I recently had a cluster where wals for the namespace table would fail lease recovery. That means ~15 minutes to make a go of recovery post-timeout, which didn't play well with the 5 minute timeout on becoming the active master. Would be nice to see that come up in integration tests.

> ChaosMonkey should inject lease recovery failures
> -------------------------------------------------
>
>                 Key: HBASE-13553
>                 URL: https://issues.apache.org/jira/browse/HBASE-13553
>             Project: HBase
>          Issue Type: New Feature
>          Components: Filesystem Integration, integration tests, MTTR
>            Reporter: Sean Busbey
>             Fix For: 2.0.0
>
>
> Update ChaosMonkey to do inject FileSystem failures around lease recovery. Would be good for seeing the impact of the long recovery timeout with other parts of the system.



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