[
https://issues.apache.org/jira/browse/SOLR-9555?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15830288#comment-15830288
]
Mark Miller commented on SOLR-9555:
-----------------------------------
bq. A replica should only ever publish its own state.
To be more precise, in normal mode of operation, this is true. You can imagine
a further improvement in line with another JIRA we have that enables a replica
to try and publish itself as DOWN if sees local resource issues. You can
imagine a case where the replica might maintain it's ZK connection but not be
operational. In this case a leader might recognize that and take hard core
action like publishing DOWN for that replica. It should only do this as an
emergency action though, after it's clear the replica still appears up but is
useless or detrimental. In this common path for LIR, we should only be trying
to communicate directly and through ZooKeeper and do no false publishing.
> Leader incorrectly publishes state for replica when it puts replica into LIR.
> -----------------------------------------------------------------------------
>
> 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]
