[
https://issues.apache.org/jira/browse/BOOKKEEPER-199?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13436681#comment-13436681
]
Rakesh R commented on BOOKKEEPER-199:
-------------------------------------
@Vinay
I am thinking of introducing one daemon thread to check the disk periodically.
We can configure the threshold. If any disk is full more than the threshold,
then that disk can be excluded from next writes.
Yeah, I also feel we can follow the similar pattern used in hdfs. If I
understand your thoughts correctly, we will do the following changes:
- When there is any IOException while adding entry, just make it simple by
shutdown the bk server(I think not required to have comlpex logic to classify
disk full or disk failure on IOE)
- 'threshold' : checks the disk percentage usage rather than plain disk size.
Disk percentage will be a configured item in bk_server.conf
- 'ResourceMonitor' daemon thread and is periodically checking disk full in
background. Inorder to control the concurrency level between the scanner and
the addentry client, we shouldn't allow entries beyond 85% or 90% of disk size.
Also scanner can look for disk error and initiate bk shutdown.
Would like to see others thoughts?
> 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: Bug
> 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
