[
https://issues.apache.org/jira/browse/CASSANDRA-18757?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17788661#comment-17788661
]
Stefan Miklosovic edited comment on CASSANDRA-18757 at 11/22/23 10:36 AM:
--------------------------------------------------------------------------
[~ethan.brown] I am not able to run a multiplexer in my CI because there is a
timeout of 1 hour in my jobs. The whole test changes you did take like 5
minutes locally and it adds up pretty quickly. (the test you rewrote just
recently). Would you mind to lower the wait there from 1 minute to 30 seconds,
for example? Is it really necessary to wait so long?
Another solution is to lower the number of the tests but there is a rule we try
to stick to that we run any (changed) test 500 times to rule out any flakiness.
I would have to split it into two runs for 250 for example which is a little
bit ... uncomfortable.
I naively tried to lower the wait from 1 minute to 30s but it fails. Maybe
there is some knob to turn to change it from 1 minute to 30 seconds? What
influences that 1 minute gap?
[https://app.circleci.com/pipelines/github/instaclustr/cassandra/3540/workflows/6b8040ec-e5a8-4f45-a346-8afd8c99a961/jobs/155512]
was (Author: smiklosovic):
[~ethan.brown] I am not able to run a multiplexer in my CI because there is a
timeout of 1 hour in my jobs. The whole test changes you did take like 5
minutes locally and it adds up pretty quickly. (the test you rewrote just
recently). Would you mind to lower the wait there from 1 minute to 30 seconds,
for example? Is it really necessary to wait so long?
Another solution is to lower the number of the tests but there is a rule we try
to stick to that we run any (changed) test 500 times to rule out any flakiness.
I would have to split it into two runs for 250 for example which is a little
bit ... uncomfortable.
[https://app.circleci.com/pipelines/github/instaclustr/cassandra/3540/workflows/6b8040ec-e5a8-4f45-a346-8afd8c99a961/jobs/155512]
> UnifiedCompactionTask is incorrectly setting keepOriginals
> ----------------------------------------------------------
>
> Key: CASSANDRA-18757
> URL: https://issues.apache.org/jira/browse/CASSANDRA-18757
> Project: Cassandra
> Issue Type: Bug
> Components: Local/Compaction
> Reporter: Branimir Lambov
> Assignee: Ethan Brown
> Priority: Normal
> Fix For: 5.0-beta, 5.x
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> {code:java}
> super(cfs, txn, gcBefore,
> strategy.getController().getIgnoreOverlapsInExpirationCheck());{code}
> in {{UnifiedCompactionTask}} is calling the base constructor
> {code:java}
> public CompactionTask(ColumnFamilyStore cfs, LifecycleTransaction txn, long
> gcBefore, boolean keepOriginals)
> {code}
> which can set {{keepOriginals}} to true when it should not be.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
