[ https://issues.apache.org/jira/browse/HBASE-3610?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Andrew Purtell updated HBASE-3610: ---------------------------------- Fix Version/s: 0.92.0 Bringing into 0.92. Reasonable for 0.90.2? > Improve RegionSplitter Performance > ---------------------------------- > > Key: HBASE-3610 > URL: https://issues.apache.org/jira/browse/HBASE-3610 > Project: HBase > Issue Type: Improvement > Reporter: Nicolas Spiegelberg > Assignee: Nicolas Spiegelberg > Priority: Trivial > Fix For: 0.92.0 > > Attachments: 0001-HBASE-3610-Improve-RegionSplitter-performance.patch > > > When running RegionSplitter on a 100-node cluster with 900 regions (and > plenty of data), the utility took around 72 hours to run. Analysis revealed > two major bottlenecks: > 1. We are serialized on the logical split (i.e. waiting for the split request > to be registered). Parallelizing this step will align configured and actual > outstanding splits. > 2. Outstanding splits are modeled like a queue. Changing this to a list with > a scanner will allow handling splits that finish out of order. -- This message is automatically generated by JIRA. For more information on JIRA, see: http://www.atlassian.com/software/jira