[
https://issues.apache.org/jira/browse/CASSANDRA-9318?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14604767#comment-14604767
]
Jonathan Ellis commented on CASSANDRA-9318:
-------------------------------------------
I think you're [Aleksey] confusing replica-side and coordinator-side. There
are no separate write and read queues on the coordinator and there probably
shouldn't be.
Again, let's avoid the temptation to boil the ocean. We can make a simple
improvement by bounding the total outstanding request size that we have in
flight on the coordinator side. Once we exceed that global limit we stop
reading additional requests. That's it.
There are lots of ways we can tweak this (e.g. estimating how many rows a read
is likely to return) but making that change alone will be a big improvement and
is something that we can reasonably do for 2.1.x, so let's not hold it up for
grand rewrites of all the things.
> Bound the number of in-flight requests at the coordinator
> ---------------------------------------------------------
>
> Key: CASSANDRA-9318
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9318
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Ariel Weisberg
> Assignee: Ariel Weisberg
> Fix For: 2.1.x, 2.2.x
>
>
> It's possible to somewhat bound the amount of load accepted into the cluster
> by bounding the number of in-flight requests and request bytes.
> An implementation might do something like track the number of outstanding
> bytes and requests and if it reaches a high watermark disable read on client
> connections until it goes back below some low watermark.
> Need to make sure that disabling read on the client connection won't
> introduce other issues.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
