[
https://issues.apache.org/jira/browse/SOLR-13386?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16818195#comment-16818195
]
Mark Miller commented on SOLR-13386:
------------------------------------
bq. I imagine this is basically impossible to write a unit test for?
The only way I've seen it spin is some very strange case and I think it
requires more than one Overseer to overlap and it was a slightly different
version of similiar code.
Trying to catch some case of it and then assert ZK was not hit too often is a
tough ask I think. We have to dig for a very uncommon case of that loop that we
may or may not be able to tingle and then everything still works so we just
have to see how many calls to ZK where made and decide if it was too much. So
yeah, I'm not sure we can add a lot of value with a test myself.
> Remove race in OverseerTaskQueue#remove that can result in the Overseer
> causing a Zookeeper call spin spike.
> ------------------------------------------------------------------------------------------------------------
>
> Key: SOLR-13386
> URL: https://issues.apache.org/jira/browse/SOLR-13386
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Mark Miller
> Assignee: Mark Miller
> Priority: Major
> Fix For: 7.7.2, 8.1
>
> Attachments: SOLR-13386.patch
>
>
> If the data call hits NoNodeException, it will throw and the Overseer work
> queue processor will catch it and loop and repeat, which causes major zk
> getData / NoNode call traffic or other such things.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
