[
https://issues.apache.org/jira/browse/CASSANDRA-21456?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sam Tunnicliffe updated CASSANDRA-21456:
----------------------------------------
Change Category: Performance
Complexity: Normal
Fix Version/s: 6.x
Reviewers: Sam Tunnicliffe
Status: Open (was: Triage Needed)
> Add policy for selecting CMS host when submitting commit request
> ----------------------------------------------------------------
>
> Key: CASSANDRA-21456
> URL: https://issues.apache.org/jira/browse/CASSANDRA-21456
> Project: Apache Cassandra
> Issue Type: Improvement
> Components: Transactional Cluster Metadata
> Reporter: Sam Tunnicliffe
> Assignee: Jon Meredith
> Priority: Normal
> Fix For: 6.x
>
>
> While running experiments on clusters with 1000+ nodes and multiple
> concurrent range movements, we found a lot of paxos contention on the CMS
> instances. Which CMS member an instance submits requests to is currently
> selected at random, causing the contention and ultimately leading to write
> timeout failures.
> One option is the addition of a policy setting to choose between random and
> deterministic CMS member selection and whether to prefer the local DC for
> changes or select globally.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
