[ https://issues.apache.org/jira/browse/HBASE-9775?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13797116#comment-13797116 ]
Nick Dimiduk commented on HBASE-9775: ------------------------------------- bq. The RS's are running G1GC so there should be no issues with GC pausing Any chance of running with old CMS settings? Let's eliminate as many variables as we can. Along those lines, is there we can isolate the impact of other major changes, such as old vs new RPC? bq. I fully agree, but do you know how many threads you have on the 0.94 during the test? we're not bounded there iirc, so as the workload is random i would expect all batch calls to go on all servers. I'm also curious about these numbers. What we should be interested in here are the active threads sending data (vs. those from 94 holding open/pending connections and in 96 which are in their RTBE retry loop) > Client write path perf issues > ----------------------------- > > Key: HBASE-9775 > URL: https://issues.apache.org/jira/browse/HBASE-9775 > Project: HBase > Issue Type: Bug > Components: Client > Affects Versions: 0.96.0 > Reporter: Elliott Clark > Priority: Critical > Attachments: Charts Search Cloudera Manager - ITBLL.png, Charts > Search Cloudera Manager.png, job_run.log, short_ycsb.png, > ycsb_insert_94_vs_96.png > > > Testing on larger clusters has not had the desired throughput increases. -- This message was sent by Atlassian JIRA (v6.1#6144)