[
https://issues.apache.org/jira/browse/TRAFODION-337?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14721184#comment-14721184
]
Suresh Subbiah commented on TRAFODION-337:
------------------------------------------
Marvin Anderson (marvin-anderson) on 2014-06-02
tags: added: data-corruption
Stacey Johnson (sjohnson-w) on 2014-06-10
information type: Proprietary → Public
Suresh Subbiah (suresh-subbiah) wrote on 2014-06-13: #1
Message from Marvin ."We have seen this problem on other than upgrades. The
most recent occurrence happened to coincide with an upgrade but we do don’t
believe upgrade is specifically the cause. We have encountered this problem
about 10 times, only once was it during an upgrade. I know there is not much to
go on with this bug and I doubt there will ever be a reproducible test case.
The best we can do is turn a system over for debugging the next time it gets
into this clobbered state."
Changed in trafodion:
assignee: nobody → Suresh Subbiah (suresh-subbiah)
Marvin Anderson (marvin-anderson) wrote on 2014-11-14: #2
This bug was entered originally to make the public aware of these random
issues. There really isn’t any more details to add, we have seen random
clobberings of tables, developers have gotten on those clusters and looked
around but found nothing specific. It’s a difficult issue to debug since no one
action can be linked to clobbering the tables. We now have the expertise to at
least wipe the tables and restart no matter how badly they are clobbered and
that’s what we do now. At the time this bug was written we lacked that
expertise.
I'm closing this bug and we'll enter a new bug (hopefully with more details) if
it occurs in the future.
Changed in trafodion:
assignee: Suresh Subbiah (suresh-subbiah) → nobody
status: New → Won't Fix
> LP Bug: 1324998 - HBase instability
> -----------------------------------
>
> Key: TRAFODION-337
> URL: https://issues.apache.org/jira/browse/TRAFODION-337
> Project: Apache Trafodion
> Issue Type: Bug
> Reporter: Marvin Anderson
> Priority: Blocker
> Labels: data-corruption
>
> This has been a random but reoccuring issue, most recently caused with
> initialize trafodion,upgrade. Hbase gets into a state where initialize
> trafodion,drop will not work. One workaround is sometimes going to the hbase
> shell and doing disable_all "*.*" and drop_all "*.*" will clean things up and
> allow initialzie trafodion to work so the cluster is usable. However,
> sometimes even those commands don't work and just hang.
> This might be an instability in HBase itself or something that Trafodion is
> causing to happen. It is very worrisome if a customer gets into this
> situation that they won't be able to recover from.
> We have seen this enough to justify entering a bug for it so people are aware.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
