hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kihwal Lee (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-3962) NN should periodically check writability of 'required' journals
Date Fri, 03 May 2013 15:42:17 GMT

    [ https://issues.apache.org/jira/browse/HDFS-3962?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13648505#comment-13648505
] 

Kihwal Lee commented on HDFS-3962:
----------------------------------

Here is one more thing I though about. I think NN exits when a required journal is unavailable.
This is fine for start-up, but during run-time something else can be tried. If the failed
one is the shared edit dir, some users may want to have an option to continue serving if FC
determines SBN is dead or experiencing the same problem. The fail-over may be disabled at
this point. This will increase the service availability.  The downside is that we cannot assume
the shared edit dir is always source of truth. But I think this will be a special case where
manual recovery and manual re-enabling of fail-over are needed.
                
> NN should periodically check writability of 'required' journals
> ---------------------------------------------------------------
>
>                 Key: HDFS-3962
>                 URL: https://issues.apache.org/jira/browse/HDFS-3962
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: ha, namenode
>    Affects Versions: 3.0.0, 2.0.1-alpha
>            Reporter: Todd Lipcon
>            Assignee: Todd Lipcon
>         Attachments: hdfs-3962.txt
>
>
> Currently, our HA design ensures "write fencing" by having the failover controller call
a fencing script before transitioning a new node to active. However, if the fencing script
is based on storage fencing (and not stonith), there is no _read_ fencing. That is to say,
the old active may continue to believe himself active for an unbounded amount of time, assuming
that it does not try to write to its edit log.
> This isn't super problematic, but it would be beneficial for monitoring, etc, to have
the old NN periodically check the writability of any "required" journals, and abort if they
become unwritable, even if there are no writes coming into the system.

--
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