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-1884) THBase WAL recovery sometimes needs to access __GLOBAL_TRX_TABLE__, but it may be stuck offline waiting on current region opening
Date Sat, 03 Oct 2009 05:13:23 GMT

    [ https://issues.apache.org/jira/browse/HBASE-1884?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12761810#action_12761810
] 

stack commented on HBASE-1884:
------------------------------

Whats _GLOBAL_TRK_LOG Clint?  I think I saw it in an non-THBase installl.  Could that be possible?
 Thanks.

> THBase WAL recovery sometimes needs to access __GLOBAL_TRX_TABLE__, but it may be stuck
offline waiting on current region opening
> ---------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-1884
>                 URL: https://issues.apache.org/jira/browse/HBASE-1884
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: contrib
>            Reporter: Clint Morgan
>
> Coming back after a crash, we have region A, and region for __GLOBAL_TRX_LOG. Assuming
that there was a pending trx in A at the time of commit, it will find that in the WAL, and
need to ask __GLOBAL_TRX_LOG__ what happend with that transaction. However, if we are opening
region A before region __GLOBAL_TRX_LOG__ then we will fail to be able to query as we are
blocking the opening of the __GLOBAL_TRX_LOG__ region, but we need it to open region A.

-- 
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