[
https://issues.apache.org/jira/browse/SOLR-5325?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13793666#comment-13793666
]
Mark Miller commented on SOLR-5325:
-----------------------------------
I think that the reason that this is hard to catch in a test is that we try and
do retries on connectionloss up to the expiration time - there must be some
case where we were still getting a connectionloss and no expiration though.
This issue should handle that case for this particular bit of code, but as an
overall precautionary measure, I have also bumped up the retries just a bit to
try and ensure they are going beyond the session timeout.
> zk connection loss causes overseer leader loss
> ----------------------------------------------
>
> Key: SOLR-5325
> URL: https://issues.apache.org/jira/browse/SOLR-5325
> Project: Solr
> Issue Type: Bug
> Affects Versions: 4.3, 4.4, 4.5
> Reporter: Christine Poerschke
> Assignee: Mark Miller
> Fix For: 4.5.1, 4.6, 5.0
>
> Attachments: SOLR-5325.patch, SOLR-5325.patch, SOLR-5325.patch
>
>
> The problem we saw was that when the solr overseer leader experienced
> temporary zk connectivity problems it stopped processing overseer queue
> events.
> This first happened when quorum within the external zk ensemble was lost due
> to too many zookeepers being stopped (similar to SOLR-5199). The second time
> it happened when there was a sufficient number of zookeepers but they were
> holding zookeeper leadership elections and thus refused connections (the
> elections were taking several seconds, we were using the default
> zookeeper.cnxTimeout=5s value and it was hit for one ensemble member).
--
This message was sent by Atlassian JIRA
(v6.1#6144)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
