hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "stack (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-1149) Dataloss when master and region server die at same time
Date Fri, 23 Jan 2009 19:11:59 GMT

    [ https://issues.apache.org/jira/browse/HBASE-1149?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12666655#action_12666655
] 

stack commented on HBASE-1149:
------------------------------

Yeah, known issue particular acute in standalone mode where master and regionserver run in
same process.  HADOOP-4379 means we lose edits still until 0.19.1 hopefully.  Also, HBASE-698
makes it so we don't recover if master is killed (this functionality is currently being worked
on as part of the zookeeper integration and should be added or hbase 0.20.0).

> Dataloss when master and region server die at same time
> -------------------------------------------------------
>
>                 Key: HBASE-1149
>                 URL: https://issues.apache.org/jira/browse/HBASE-1149
>             Project: Hadoop HBase
>          Issue Type: Bug
>    Affects Versions: 0.19.0
>            Reporter: Ben Maurer
>
> To reproduce:
> 1) Run HBase in standalone mode
> 2)
> create 'foo', 'bar'
> 3) kill -9 the HBase server
> 4) Restart hbase
> The table 'foo' will not exist.
> Apparently this problem happens because the master and region servers die at the same
time. To me that suggests a fairly large flaw -- if your cluster has a systematic failure
(say, a power outage) it would cause data loss.

-- 
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