[
https://issues.apache.org/jira/browse/CASSANDRA-3578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13817147#comment-13817147
]
Benedict commented on CASSANDRA-3578:
-------------------------------------
bq. We don't guarantee that in PeriodicCommitLogExecutorService
No, you're right. I misspoke. For PeriodicCLE we only don't *commit* anything
we cannot later replay.
Unfortunately the BatchCLE in your patch could still ACK messages that weren't
replayable, as a message could reserve space, not finish writing it, and so not
get placed on the waitIfNeeded() call/queue, but a later writer could race past
it and get there. Since the earlier writer hasn't serialized its size, the
later message is not replayable. Besides, the queue has *two* locks, so you'd
still have a lock ;-)
> Multithreaded commitlog
> -----------------------
>
> Key: CASSANDRA-3578
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3578
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Jonathan Ellis
> Assignee: Vijay
> Priority: Minor
> Labels: performance
> Attachments: 0001-CASSANDRA-3578.patch, ComitlogStress.java,
> Current-CL.png, Multi-Threded-CL.png, parallel_commit_log_2.patch
>
>
> Brian Aker pointed out a while ago that allowing multiple threads to modify
> the commitlog simultaneously (reserving space for each with a CAS first, the
> way we do in the SlabAllocator.Region.allocate) can improve performance,
> since you're not bottlenecking on a single thread to do all the copying and
> CRC computation.
> Now that we use mmap'd CommitLog segments (CASSANDRA-3411) this becomes
> doable.
> (moved from CASSANDRA-622, which was getting a bit muddled.)
--
This message was sent by Atlassian JIRA
(v6.1#6144)
