[ https://issues.apache.org/jira/browse/SOLR-9555?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15980740#comment-15980740 ]
Mark Miller commented on SOLR-9555: ----------------------------------- I commented a bit more about this in SOLR-10525, but the gist is, I think we should take this opportunity to stop trying to do LIR per update. We don't want update volume to be in control of how much chatter this generates any longer. What we really need is to simply trigger something that says, let the replica know it needs to enter recovery, and send that to ZK at some throttled rate that has a chatter / delay tradeoff that we control. Even if we allowed this once or twice per second or something, thats drastically better than the 1000's per second you can hit now. > 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 > Attachments: SOLR-9555.patch, SOLR-9555-WIP-2.patch, > SOLR-9555-WIP-3.patch, SOLR-9555-WIP.patch > > > See > https://jenkins.thetaphi.de/job/Lucene-Solr-master-Linux/17888/consoleFull > for an example -- This message was sent by Atlassian JIRA (v6.3.15#6346) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org