[
https://issues.apache.org/jira/browse/BOOKKEEPER-199?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13266513#comment-13266513
]
Uma Maheswara Rao G commented on BOOKKEEPER-199:
------------------------------------------------
Yes, my point is that, when we try to save the edits in Namenode dirs (lets say
4 volumes), if all 4 volumes failed, we will sutdown the node. Because this is
a serious problem. Here also we can make it read only. But it is important to
make it noticeable to admins.
Same way, Why can't we make the bookie down when all ledger disks failed. So,
that admin can take action immediateltly because one node down in cluster and
it is very much noticeable. *We can provide read only mode as option to start
the node in that mode, instead of converting automatically.* If admin realizes
that, he don't have any other option to replace the node or con't correct, that
will user headache to start the node with read only option.
If we silently convert to read-only mode, user may not get any failures. So,
he may think that cluster is fine. But one node might have been converted to
read only mode and not able to store the new data any more.
What I am trying to say is, make the serious failures noticeable and make it
fail loudly.
Am I missing some thing here?
> Provide bookie readonly mode, when journal/ledgers flushing has failed with
> IOE
> -------------------------------------------------------------------------------
>
> Key: BOOKKEEPER-199
> URL: https://issues.apache.org/jira/browse/BOOKKEEPER-199
> Project: Bookkeeper
> Issue Type: Sub-task
> Components: bookkeeper-server
> Affects Versions: 4.0.0
> Reporter: Rakesh R
> Assignee: Rakesh R
> Priority: Critical
> Fix For: 4.2.0
>
> Attachments: BOOKKEEPER-199.patch
>
>
> Bookkeeper should change to readonly(r-o) mode when the journal/ledgers
> flushing has failed with IOException. Later on, reject write requests on
> server side and will accept only the read requests from the clients, because
> even if flushing fails, the data in the bookie which has been flushed is
> still valid.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
