[
https://issues.apache.org/jira/browse/BOOKKEEPER-199?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13254635#comment-13254635
]
Sijie Guo commented on BOOKKEEPER-199:
--------------------------------------
{quote}
I am keeping the r-o status in bookie's presence znode(which is emphemeral),
when restarting its creating new presence znode and like a normal r-w bookie.
{quote}
3 bookies, b1, b2, b3.
at time t1, BookieWatcher has
availableBookies: b1, b2, b3,
roBookies: empty
at time t2, bookie b3 is turned to R-O mode. BookieWatcher changes the lists to:
availableBookies: b1, b2, b3
roBookies: b3
#getAdditionalBookie will get bookie from (b1,b2)
at time t3, bookie b3 is restarted to R-O mode. BookieWatcher's list is still
same as t2.
since b3 will not be removed from roBookies list.
so #getAdditionalBookie will still just can get bookies from (b1, b2), right?
{quote}
Hope you are pointing me to consider the ledger failures of 'SyncThread' ?
Also, I missed dealing the ledgerCache failures, it would be great to see your
thoughts/opinion ?
{quote}
SyncThread is one place.
The other place is in LedgerCache, it would call #flushLedger when evicting
pages.
BTW, from the title of this jira, it also need to handle journal flushing
failure, right? maybe you need to change code in Bookie#run() to turn a bookie
into a R-O mode instead of shutting down it directly.
> 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: 4.1.0
>
> 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