[ https://issues.apache.org/jira/browse/SOLR-6816?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14240312#comment-14240312 ]
Timothy Potter commented on SOLR-6816: -------------------------------------- I think messing with the queue poll time in CUSS is a red herring for improving indexing performance. It does help reduce CPU load on replicas (less requests / garbage) so might still be a config property we want to add, but I wasn't able to get it to always be faster than using 0, which I suppose is a good thing since that's what we're doing now in released code. The biggest performance gain I've been able to make (in addition to increasing the number of version buckets), is to avoid the lookup to the ulog (in the replica) when doing bulk adds. Of course, I realize that is special case, but doing high-volume bulk adds is when you especially need better indexing performance. > Review SolrCloud Indexing Performance. > -------------------------------------- > > Key: SOLR-6816 > URL: https://issues.apache.org/jira/browse/SOLR-6816 > Project: Solr > Issue Type: Task > Components: SolrCloud > Reporter: Mark Miller > Priority: Critical > Attachments: SolrBench.pdf > > > We have never really focused on indexing performance, just correctness and > low hanging fruit. We need to vet the performance and try to address any > holes. > Note: A common report is that adding any replication is very slow. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org