hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "HBase Review Board (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-1660) need a rolling restart script
Date Thu, 12 Aug 2010 19:48:16 GMT

    [ https://issues.apache.org/jira/browse/HBASE-1660?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12897920#action_12897920
] 

HBase Review Board commented on HBASE-1660:
-------------------------------------------

Message from: "Nicolas" <nspiegelberg@facebook.com>

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://review.cloudera.org/r/632/
-----------------------------------------------------------

(Updated 2010-08-12 12:36:09.868512)


Review request for hbase.


Changes
-------

- regionservers were being restarted in parallel.  need to restart serially
- wait for the master znode to expire before restarting the masters (this is especially necessary
with backup masters)


Summary
-------

1. added 'restart' option to hbase-daemon.sh for unit-level restarting
2. added rolling-restart.sh script to perform system-level rolling restarts

(Note that I intentionally did not restart the ZooKeeper nodes, since those binaries will
need an update far less often)


This addresses bug HBASE-1660.
    http://issues.apache.org/jira/browse/HBASE-1660


Diffs (updated)
-----

  trunk/bin/hbase-daemon.sh 984635 
  trunk/bin/regionservers.sh 984635 
  trunk/bin/rolling-restart.sh PRE-CREATION 

Diff: http://review.cloudera.org/r/632/diff


Testing
-------

./bin/start-hbase.sh
./bin/rolling-restart.sh
./bin/stop-hbase.sh


Thanks,

Nicolas




> need a rolling restart script
> -----------------------------
>
>                 Key: HBASE-1660
>                 URL: https://issues.apache.org/jira/browse/HBASE-1660
>             Project: HBase
>          Issue Type: New Feature
>    Affects Versions: 0.20.0
>            Reporter: ryan rawson
>            Priority: Minor
>             Fix For: 0.92.0
>
>
> need a script that will do a rolling restart.
> It should be configurable in 2 ways:
> - how long to keep the daemon down per host
> - how long to wait between hosts
> for regionservers in my own hacky command line I used 10/60.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message