[
https://issues.apache.org/jira/browse/ACCUMULO-1088?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13858092#comment-13858092
]
Eric Newton commented on ACCUMULO-1088:
---------------------------------------
I have a vague recollection of a change that [~kturner] made for 1.5 that made
this problem go away. I've not chased it down to a particular commit, but it
had something to do with re-establishing connections to tservers in the master
whenever there was an IOException.
> master not balancing because balance information is out-of-date
> ---------------------------------------------------------------
>
> Key: ACCUMULO-1088
> URL: https://issues.apache.org/jira/browse/ACCUMULO-1088
> Project: Accumulo
> Issue Type: Bug
> Components: master
> Affects Versions: 1.4.2
> Reporter: Eric Newton
> Assignee: Eric Newton
> Fix For: 1.5.0
>
>
> During tests with agitation, the test cluster would become unbalanced because
> the master did not realize a dead server was restarted. All servers would be
> up, but the master would think that a dead server was still out there.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
