[
https://issues.apache.org/jira/browse/BOOKKEEPER-564?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13637845#comment-13637845
]
Ivan Kelly commented on BOOKKEEPER-564:
---------------------------------------
{quote}
if you could do the optimization without ledger storage offering the hint, show
the code.
{quote}
I'm not sure which exact optimization you're referring to. My patch from the
9th of April has all the things I've discussed except for periodic syncs,
which, as I understand it, we don't want anymore. To summarize what is in the
patch.
* LedgerStorage decides when to checkpoint, and runs it autonomously
* The LogMark storage is managed by LedgerStorage
* The bookie coordinates LogMark usage between the Journal and LedgerStorage.
* Both components can now be tested 100% independently. Journal doesn't even
need the ledger storage dirs anymore.
{quote}
ledger storage and journal are just bookie's components, the coordinate thing
and semantic guarantee is better to be done in bookie itself.
{quote}
And my patch conveys this. The bookie owns the journal and the ledger storage.
And controls both. The journal doesn't control the ledger storage. The ledger
storage doesn't control the journal. The ledger storage doesn't control the
bookie. The journal doesn't control the bookie. The control flow follows the
ownership. Bookie controls ledger storage and the journal, and that's it.
> Better checkpoint mechanism
> ---------------------------
>
> Key: BOOKKEEPER-564
> URL: https://issues.apache.org/jira/browse/BOOKKEEPER-564
> Project: Bookkeeper
> Issue Type: Improvement
> Components: bookkeeper-server
> Reporter: Sijie Guo
> Assignee: Sijie Guo
> Fix For: 4.3.0
>
> Attachments: 0001-BOOKKEEPER-564-Better-checkpoint-mechanism.patch,
> 0001-BOOKKEEPER-564-Better-checkpoint-mechanism.patch,
> 0002-BOOKKEEPER-564-Better-checkpoint-mechanism.patch, BOOKKEEPER-564.patch,
> BOOKKEEPER-564.patch
>
>
> Currently, SyncThread made a checkpoint too frequently, which affects
> performance. data is writing to entry logger file might be blocked by syncing
> same entry logger file, which affect bookie to achieve higher throughput. We
> could schedule checkpoint only when rotating an entry log file. so new
> incoming entries would be written to newer entry log file and old entry log
> file could be synced.
--
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
