[
https://issues.apache.org/jira/browse/CASSANDRA-21185?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18063241#comment-18063241
]
Sam Tunnicliffe commented on CASSANDRA-21185:
---------------------------------------------
ah yes, good catch. Previously, in the face of a token collision node startup
would be allow to proceed "silently" (I think it would maybe log something, but
it was definitely unsafe behaviour), hence these tests require a sequential
startup.
In trunk that can't happen as the token assignments are linearized, and C* now
behaves safely so I'd be a bit reluctant to paper over that by automatically
re-allocating and retrying. Maybe it is the right thing to do but I think it
warrants a bit of discussion (maybe on the mailing list) before making a
change.
So in the first instance I'd be inclined to go back to the original plan of
forcing sequential startup regardless of version, sound reasonable?
> Fix flaky DTest: bootstrap_test_*
> ---------------------------------
>
> Key: CASSANDRA-21185
> URL: https://issues.apache.org/jira/browse/CASSANDRA-21185
> Project: Apache Cassandra
> Issue Type: Bug
> Components: Test/dtest/python
> Reporter: Sam Lightfoot
> Assignee: Sam Lightfoot
> Priority: Normal
> Fix For: 5.1
>
> Attachments: split_brain_logs.txt
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Tests often failing due to no seed no being up whilst non-seeds trying to
> join the ring. Likely fix to start the seed node to ensure CMS initialization
> is complete then allow other nodes in CCM to start in parallel.
> Affects 5.1+ due to <=5.0 using [sequential
> startup|https://github.com/apache/cassandra-dtest/blob/trunk/bootstrap_test.py#L254C30-L254C32].
> On further analysis it appears two separate clusters form due to the seed
> node not accepting messages during CMS initialization. Attached logs show the
> independent clusters resulting from
> _bootstrap_test.py::TestBootstrap::test_read_from_bootstrapped_node._
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
