hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "stack (JIRA)" <j...@apache.org>
Subject [jira] Updated: (HBASE-3318) Split rollback leaves parent with writesEnabled=false
Date Fri, 10 Dec 2010 04:57:01 GMT

     [ https://issues.apache.org/jira/browse/HBASE-3318?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

stack updated HBASE-3318:
-------------------------

    Fix Version/s:     (was: 0.90.1)
                       (was: 0.92.0)
                   0.90.0

Bringing into 0.90.0.  Want to apply J-D?

> Split rollback leaves parent with writesEnabled=false
> -----------------------------------------------------
>
>                 Key: HBASE-3318
>                 URL: https://issues.apache.org/jira/browse/HBASE-3318
>             Project: HBase
>          Issue Type: Bug
>            Reporter: Jean-Daniel Cryans
>            Assignee: Jean-Daniel Cryans
>            Priority: Critical
>             Fix For: 0.90.0
>
>         Attachments: HBASE-3318.patch
>
>
> I saw a split rollback today, and it left the region in a state where it was able to
take writes, but wasn't able to flush or compact. It's printing this message every few milliseconds:
> {noformat}
> NOT flushing memstore for region xxx., flushing=false, writesEnabled=false
> {noformat}
> I see why, writesEnabled is never set back in HRegion.initialize:
> {code}
> // See if region is meant to run read-only.
> if (this.regionInfo.getTableDesc().isReadOnly()) {
>   this.writestate.setReadOnly(true);
> }
> {code}
> Instead it needs to pass isReadOnly into the setReadOnly method to work correctly.
> I think it should go in 0.90.0 if there's a new RC.

-- 
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