[
https://issues.apache.org/jira/browse/CASSANDRA-9318?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14558016#comment-14558016
]
Benedict commented on CASSANDRA-9318:
-------------------------------------
Perhaps we can do something a bit more constrained here: we can bound the
number of requests we will have in flight _from any single client_. Bulk
loading can then happen to a constrained number of servers, and we can avoid
even throwing OE, and instead temporarily suspend the consumption of the TCP
stream for that client (using TCP backpressure to more gracefully inform the
client to stop spamming us). This would require that the clients people are
using not use unbounded queues to store their send queue, but let's hope they
do that anyway.
> 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: 2.2.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)
