[
https://issues.apache.org/jira/browse/SOLR-8238?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18082747#comment-18082747
]
David Smiley commented on SOLR-8238:
------------------------------------
Suggested approach:
* a core loads... at some point where it's ready to receive requests...
* asynchronously dispatch a command to node level executor of only 1 thread ...
* Runnable logic:
condition 1: we (this replica) are in state "ACTIVE" (we should already be but
check just in case)
condition 2: are we *not* the leader yet we are preferredLeader? continue...
condition 3: Is the node holding the current leader a version of Solr greater
than or equal to the current node? (avoids leaders upgrading too soon)
continue...
Action: tell the current leader to give up leadership. Wait for the
leadership transition to complete so that we don't prematurely free up a thread
that could lead to many simultaneous index replications. There's a
cluster/http command to give up leadership; no ZK surgery. The effect of this
will be the current replica will become leader since a preferredLeader joins
the election "at head" (thus next in line) instead of last.
Maybe this should be registered to occur just-in-time like a state watcher /
event listener. That's probably more elegant.
> Make Solr respect preferredLeader at startup
> --------------------------------------------
>
> Key: SOLR-8238
> URL: https://issues.apache.org/jira/browse/SOLR-8238
> Project: Solr
> Issue Type: Improvement
> Components: SolrCloud
> Affects Versions: 5.2.1
> Reporter: Peter Morgan
> Priority: Minor
>
> After setting preferredLeader property, noticed that upon restarting leaders
> revert to wherever they were previously running before REBALANCE was called.
> I would expect the preferredLeader to influence the startup election, but it
> appears it is not observed.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
