[ https://issues.apache.org/jira/browse/HBASE-19344?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16266298#comment-16266298 ]
Chance Li commented on HBASE-19344: ----------------------------------- bq. And I haven’t understand the second question on sync fail, could you please explain a bit more? only suggestion, sir. When Fanout do #flush failed, AsyncFSWal will do #syncFailed putting the edit from #unackedAppends to #toWriteAppends, and rolling to new log, then redo all. maybe retry again and again util timeout by #blockOnSync. Maybe it's not enough for the AsyncWal. On this case which NN is ok, but DN has some exception, I think this can lead to create many many logs. And we can't enumerate all the exception. So maybe it's neccessary for 'define the bound'. > improve asyncWAL by using Independent thread for netty #IO in > FanOutOneBlockAsyncDFSOutput > ------------------------------------------------------------------------------------------- > > Key: HBASE-19344 > URL: https://issues.apache.org/jira/browse/HBASE-19344 > Project: HBase > Issue Type: Sub-task > Components: wal > Affects Versions: 2.0.0-beta-1 > Reporter: Chance Li > Assignee: Chance Li > Fix For: 2.0.0 > > Attachments: HBASE-19344-branch2.patch, > HBASE-19344-branch2.patch.2.POC, wal-1-test-result.png, > wal-8-test-result.png, ycsb_result_apache20_async_wal.pdf > > > The logic now is that the netty #IO thread and asyncWal's thread are the same > one. > Improvement proposal: > 1, Split into two. > 2, All multiWal share the netty #IO thread pool. -- This message was sent by Atlassian JIRA (v6.4.14#64029)