[ https://issues.apache.org/jira/browse/SOLR-11776?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Cassandra Targett updated SOLR-11776: ------------------------------------- Component/s: SolrCloud > HttpSolrClient should handle SocketException:broken pipe > -------------------------------------------------------- > > Key: SOLR-11776 > URL: https://issues.apache.org/jira/browse/SOLR-11776 > Project: Solr > Issue Type: Bug > Security Level: Public(Default Security Level. Issues are Public) > Components: SolrCloud > Reporter: Cao Manh Dat > Priority: Major > > There is a case when leader sends an update to replica but met > {{SocketException: broken pipe}} ( HttpPartitionTest ), the leader treat this > exception as others SocketException and put replica into LIR mode. But this > exception means an existing connection between leader and replica is closed > on one side. I think we should handle this exception in a different way ( ex: > reconnect ) so leader won't blindly put replica into LIR even when the > connection between leader and replica is healthy. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org