hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sergey Shelukhin (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-7268) correct local region location cache information can be overwritten w/stale information from an old server
Date Thu, 13 Dec 2012 18:38:13 GMT

    [ https://issues.apache.org/jira/browse/HBASE-7268?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13531302#comment-13531302
] 

Sergey Shelukhin commented on HBASE-7268:
-----------------------------------------

bq. How would you use it Sergey? Would it be ok if on cluster restart, the sequence restarted
at zero?
The single-source, increasing timer would be useful for any coordination tasks... e.g. you'd
always know which events happened earlier, across master restarts/etc. It should only reset
when there's a singularity e.g. if you wipe the cluster.
It's overkill to do it just for this issue though...
I think I saw it discussed somewhere, maybe in a JIRA related to snapshots.

bq. Yes. But we can't have client register to get callbacks when regions moves. What you thinking?
First sleep then get location? :)
                
> correct local region location cache information can be overwritten w/stale information
from an old server
> ---------------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-7268
>                 URL: https://issues.apache.org/jira/browse/HBASE-7268
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 0.96.0
>            Reporter: Sergey Shelukhin
>            Assignee: Sergey Shelukhin
>            Priority: Minor
>             Fix For: 0.96.0
>
>         Attachments: HBASE-7268-v0.patch, HBASE-7268-v0.patch, HBASE-7268-v1.patch, HBASE-7268-v2.patch
>
>
> Discovered via HBASE-7250; related to HBASE-5877.
> Test is writing from multiple threads.
> Server A has region R; client knows that.
> R gets moved from A to server B.
> B gets killed.
> R gets moved by master to server C.
> ~15 seconds later, client tries to write to it (on A?).
> Multiple client threads report from RegionMoved exception processing logic "R moved from
C to B", even though such transition never happened (neither in nor before the sequence described
below). Not quite sure how the client learned of the transition to C, I assume it's from meta
from some other thread...
> Then, put fails (it may fail due to accumulated errors that are not logged, which I am
investigating... but the bogus cache update is there nonwithstanding).
> I have a patch but not sure if it works, test still fails locally for yet unknown reason.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message