[ https://issues.apache.org/jira/browse/SOLR-4926?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13688339#comment-13688339 ]
Mark Miller commented on SOLR-4926: ----------------------------------- bq. the use of CFS somehow causes replication to fail Yeah, this is what I'm seeing - I just caught a really good sample case with decent logging. The recovering replica commits on the leader and that leader then has 126 docs to replicate. 16 documents end up on the relica after the replication - 110 short. The leader is on gen 3, the replica on gen 1. Perhaps a red herring, but in the many cases of this I've looked at, oddly, no buffered docs are ever replayed after that - though I have seen buffered docs replayed in those same runs when the replication did not fail. Weird observation. Anyway, I need to turn on more replication level logging I think. > I am seeing RecoveryZkTest and ChaosMonkeySafeLeaderTest fail often on trunk. > ----------------------------------------------------------------------------- > > Key: SOLR-4926 > URL: https://issues.apache.org/jira/browse/SOLR-4926 > Project: Solr > Issue Type: Bug > Components: SolrCloud > Reporter: Mark Miller > Assignee: Mark Miller > Priority: Blocker > Fix For: 5.0, 4.4 > > -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org