hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Phabricator (Commented) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-4742) Split dead server's log in parallel
Date Sat, 05 Nov 2011 19:22:52 GMT

    [ https://issues.apache.org/jira/browse/HBASE-4742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13144803#comment-13144803
] 

Phabricator commented on HBASE-4742:
------------------------------------

mbautin 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:333 My concern is
that there might be two threads at some point splitting logs for the same server.
  src/main/java/org/apache/hadoop/hbase/master/ProcessServerShutdown.java:337 "Succeed" is
infinitive and "succeeded" is past, so e.g. "succeeded to split" would be more appropriate.
But this does not really matter.
  src/main/java/org/apache/hadoop/hbase/master/ProcessServerShutdown.java:347-348 OK, this
is fine. In general, unchecked exceptions should only be thrown when the program is not expected
to recover from the error (http://bit.ly/RuntimeException). This is probably appropriate in
case of an unknown enum element.
  src/test/java/org/apache/hadoop/hbase/master/TestMultiRegionServerShutDown.java:135 What
do you mean by "this exception" and what's the purpose of ignoring it? This will ignore all
exceptions and make the code hard to debug in case something goes wrong in the try block.
  src/main/java/org/apache/hadoop/hbase/master/ProcessServerShutdown.java:312 OK, if we believe
there will be no more than say 10% of servers going down at the same time, and the current
cluster size is 100 nodes, this is fine. However, if we had 1000 servers and half of them
went down, then the master might choke trying to split 500 logs at once.

  Prakash: do we use distributed log splitting only on master startup or during normal master
operation as well?

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
>
>
> 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

        

Mime
View raw message