[ https://issues.apache.org/jira/browse/HBASE-4742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13146675#comment-13146675 ]
Phabricator commented on HBASE-4742: ------------------------------------ Liyin has commented on the revision "[jira] [HBASE-4742] Split dead server's log in parallel". INLINE COMMENTS src/main/java/org/apache/hadoop/hbase/master/ProcessServerShutdown.java:323 Hi Kannan, When testing this revision internally, I have already made this change to add this operation back to delay queue but in a different way. There are 2 ways to add this operation back to the delay queue instead of to do queue. 1) call this.requeue(), which is add the operation to the delay queue and return true. It means this operation has succeeded and all the post-process-listener will be called. 2) Throw out some exceptions here, which will be caught in the RegionServerOperationQueue. Then this operation will be put back to the delay queue bu RegionServerOperationQueue. After discussion with Prakash and Mikhail, Prakash prefers the 2nd solutions and Mikhail suggests to refactor the code. So each operation should return OPERATION.SUCCESS, OPERATION.FAILED and OPERATION.DLAYED. I have already implemented the 2nd solutions internally and get it tested in big cluster. But now I try to refactor the code as Mikhail suggested. I will update this revision once every thing settled down during the test. REVISION DETAIL https://reviews.facebook.net/D237 > Split dead server's log in parallel > ----------------------------------- > > Key: HBASE-4742 > URL: https://issues.apache.org/jira/browse/HBASE-4742 > Project: HBase > Issue Type: Improvement > Reporter: Liyin Tang > Assignee: Liyin Tang > Attachments: D237.1.patch, D237.2.patch, D237.3.patch, D237.4.patch, > D237.5.patch, D237.6.patch, D237.7.patch, D237.8.patch > > > When one region server goes down, the master will shutdown the region server > and split its log. > However, splitting log is a blocking call and it would take some time. > If more than one region server go down, the master will split its log one by > one, which is not efficient. > Since we have the distributed log split, we could split these logs from the > dead servers in parallel. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira