[
https://issues.apache.org/jira/browse/CASSANDRA-19050?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17790463#comment-17790463
]
Alex Petrov edited comment on CASSANDRA-19050 at 11/28/23 9:25 AM:
-------------------------------------------------------------------
Wanted to mention that TCM merge related tests are
[documented|https://issues.apache.org/jira/browse/CASSANDRA-19055], and there
are rather few of them by now. If you see any failure that is not in the list,
there’s a chance it is not TCM related. There have been flakes we have fixed
that existed in trunk, too.
I suggest in this interim period, we list test failures and cross-reference
them against the TCM list. For example,
org.apache.cassandra.index.sai.cql.StorageAttachedIndexDDLTest#concurrentTruncateWithIndexBuilding-oa_jdk17
does not seem to be TCM related, and there is an issue open about it:
[CASSANDRA-19032]. Another that looks like unfamiliar one is
https://app.circleci.com/pipelines/github/bereng/cassandra/1122/workflows/31c7e520-ca04-4278-b0c8-a9fa32b3a0f9/jobs/34616/tests
I have not checked all pipelines, but this is something that stood out. Does
not seem to be broken because of this patch though.
I just think attributing all test failures, especially flakes, to TCM is also
not the best way to go forward.
was (Author: ifesdjeen):
Wanted to mention that TCM merge related tests are
[documented|https://issues.apache.org/jira/browse/CASSANDRA-19055], and there
are rather few of them by now. If you see any failure that is not in the list,
there’s a chance it is not TCM related. There have been flakes we have fixed
that existed in trunk, too.
I suggest in this interim period, we list test failures and cross-reference
them against the TCM list. For example,
org.apache.cassandra.index.sai.cql.StorageAttachedIndexDDLTest#concurrentTruncateWithIndexBuilding-oa_jdk17
is not TCM related, and there seems to be an issue open about it:
[CASSANDRA-19032]. Another that looks like unfamiliar one is
https://app.circleci.com/pipelines/github/bereng/cassandra/1122/workflows/31c7e520-ca04-4278-b0c8-a9fa32b3a0f9/jobs/34616/tests
I have not checked all pipelines, but this is something that stood out. Does
not seem to be broken because of this patch though.
I just think attributing all test failures, especially flakes, to TCM is also
not the best way to go forward.
> Enhanced usage of test method names in CQLTester for better test debugging
> --------------------------------------------------------------------------
>
> Key: CASSANDRA-19050
> URL: https://issues.apache.org/jira/browse/CASSANDRA-19050
> Project: Cassandra
> Issue Type: Improvement
> Components: Test/unit
> Reporter: Berenguer Blasi
> Assignee: Berenguer Blasi
> Priority: Normal
> Fix For: 5.0-rc, 5.x
>
>
> All views, tables, ks,... created in CQLTester are of the form {{table_00}}:
> - While debugging tests and flakies this makes it really hard to match logs
> to test methods.
> - Some async operations can spill log lines from a previous test method into
> the next's log lines which is incredibly confusing
> - It's hard on the eyes and easy to mix up
> - When comparing logs from 2 different branches, envs,... it's really hard to
> match logs
> The proposed solution is for {{CQLTester}} to decorate these with the test
> name.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
