[
https://issues.apache.org/jira/browse/CASSANDRA-19169?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17793815#comment-17793815
]
Sam Tunnicliffe commented on CASSANDRA-19169:
---------------------------------------------
Trivial patch
[here|https://github.com/beobal/cassandra/commit/cd456f7e30f6128e67631503f0e71f44444b99cc]
Circle CI runs for branch with CASSANDRA-19169, CASSANDRA-19102 &
CASSANDRA-19171:
[J11|https://app.circleci.com/pipelines/github/michaelsembwever/cassandra/274/workflows/6843aabd-4749-4cbf-94a5-ec3a546704e6],
[J17|https://app.circleci.com/pipelines/github/michaelsembwever/cassandra/274/workflows/440e2e39-038e-45ba-ab91-80335a497f21]
Aside from some strangeness with cqlshlib tests which appears completely
unrelated, only failures are the ones tracked in CASSANDRA-19072,
CASSANDRA-19058 & CASSANDRA-18360
> Don't NPE when initializing CFSs for local system keyspaces with UCS
> --------------------------------------------------------------------
>
> Key: CASSANDRA-19169
> URL: https://issues.apache.org/jira/browse/CASSANDRA-19169
> Project: Cassandra
> Issue Type: Bug
> Components: Transactional Cluster Metadata
> Reporter: Sam Tunnicliffe
> Assignee: Sam Tunnicliffe
> Priority: Normal
> Fix For: 5.1-alpha1
>
>
> When UnifiedCompactionStrategy is used as the default, NPEs are thrown when
> flushing the system keyspace tables early during startup. The system keyspace
> is
> initialised before the cluster metadata, but UCS currently tries to access the
> current epoch when initialising the shard manager, to determine whether the
> local ranges are out of date. This isn't necessary for the system keyspaces as
> they use LocalStrategy and cover the whole token space.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
