[
https://issues.apache.org/jira/browse/CASSANDRA-3578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13834803#comment-13834803
]
Benedict commented on CASSANDRA-3578:
-------------------------------------
Thinking a little more on the PCLE guarantees about syncing, I think perhaps a
better scheme is to attempt to sync() twice every "poll interval" (or once
every 0.49*interval), and to only update the heartbeat at the end of a sync();
at the moment I update before and after, to deal with the fact that the
blockInterval and pollInterval are approximately the same (so we don't simply
block every time we start a sync()). This means you could lose upto 2x interval
commits. I addressed this in a previous comment, but didn't deal with it in the
code.
It also occurs to me that pollInterval is a really bad name, and it should be
called maxSyncInterval or something.
> Multithreaded commitlog
> -----------------------
>
> Key: CASSANDRA-3578
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3578
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Jonathan Ellis
> Assignee: Benedict
> Priority: Minor
> Labels: performance
> Attachments: 0001-CASSANDRA-3578.patch, ComitlogStress.java,
> Current-CL.png, Multi-Threded-CL.png, latency.svg, oprate.svg,
> 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)
