[jira] [Updated] (SOLR-5833) OVERSEEROP aka ZkController.forceOverSeer doesn't work: old overser keeps running

["Cassandra Targett (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SOLR-5833?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Cassandra Targett updated SOLR-5833:
------------------------------------
    Component/s: SolrCloud

> OVERSEEROP aka ZkController.forceOverSeer doesn't work: old overser keeps running
> ---------------------------------------------------------------------------------
>
>                 Key: SOLR-5833
>                 URL: https://issues.apache.org/jira/browse/SOLR-5833
>             Project: Solr
>          Issue Type: Bug
>          Components: SolrCloud
>            Reporter: Hoss Man
>
> While working on SOLR-5823, i tried to write a test using ZkController.forceOverSeer() to force the overseer to change from one node to another (to prove my code could properly detect the new overseer)
> The test failed spectacuarly, because after calling forceOverSeer() the end result was that Overseer obejcts on both ZkControllers wound up being active.
> I'm not well versed on the leader/overseer election stuff -- but from what i can tell poking arround in the code of forceOverSeer(), compared to the code involved in OversearchElectionContext, it seems like forceOverSeer() doesn't relaly do anything to ensure that the "old" overseer ever "stops".  miller's comments in SOLR-5823 seem to re-affirm that.
> We should make forceOverSeer() work safely, or we should eliminate it.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org