[
https://issues.apache.org/jira/browse/CASSANDRA-21072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18054199#comment-18054199
]
Arnav Chakraborty commented on CASSANDRA-21072:
-----------------------------------------------
Hi [~brandon.williams] ,
I see we're only handling TombstoneOverwhelmingException in the attached
datastax commit, but there are other RejectException subclasses like
LocalReadSizeTooLargeException, QueryReferencingTooManyIndexesException, etc.
Shouldn't ALL RejectExceptions propagate to the client when warnings tracking
is enabled? They're all guardrail violations.
Would it be cleaner to either:
1. Remove this catch block entirely (let all RejectExceptions propagate), or
2. Check instanceof RejectException to handle all subclasses?
Also can I assign this task to myself , I am new to Cassandra and want to begin
contributing to the project.
> TombstoneOverwhelmingException gets swallowed when tracking warnings
> --------------------------------------------------------------------
>
> Key: CASSANDRA-21072
> URL: https://issues.apache.org/jira/browse/CASSANDRA-21072
> Project: Apache Cassandra
> Issue Type: Bug
> Components: Feature/Guardrails
> Reporter: Brandon Williams
> Priority: Normal
> Fix For: 5.0.x, 6.x
>
>
> ReadCommandVerbHandler is catching RejectException and returning an empty
> response instead of propagating the failure when warnings tracking is
> enabled, see
> https://github.com/datastax/cassandra/commit/f1223ec9fd9c724f00acbe338e053f3533617c8a
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
