[
https://issues.apache.org/jira/browse/SOLR-9555?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15827656#comment-15827656
]
Cao Manh Dat commented on SOLR-9555:
------------------------------------
This issue seems related to SOLR-9945.
As I discussed with [~shalinmangar],
In SOLR-9945, the reason that leader can not send the update to replica because
the underlying socket is closed on replica side, so the leader will put replica
into LIR state.
I don't think we should fix the problem by "only set DOWN if the current state
is ACTIVE" , because we did not send the update to the replica, so the replica
won't have that update, we should put the replica into LIR, it's a right thing
to do. I think we should fix the problem by "close" the underlying socket based
on ZK events.
> Recovery can hang if a node is put into LIR as it is starting up
> ----------------------------------------------------------------
>
> Key: SOLR-9555
> URL: https://issues.apache.org/jira/browse/SOLR-9555
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Alan Woodward
>
> See
> https://jenkins.thetaphi.de/job/Lucene-Solr-master-Linux/17888/consoleFull
> for an example
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
