[
https://issues.apache.org/jira/browse/CASSANDRA-17729?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17562954#comment-17562954
]
Berenguer Blasi commented on CASSANDRA-17729:
---------------------------------------------
^Ah well yes apologies. It works better if you actually link sthg to it :-)
[~jmckenzie] mentioned on the ML:
{quote}Another option would be to increase the resources dedicated to each
agent container and run less in parallel. Or, best yet, do both (up timeouts
and lower parallelization / up resources).
As far as I can tell the failures on Jenkins aren't value-add compared to what
we're seeing on circleci and are just generating busywork.
There's a reasonable discussion to be had about "what's the smallest footprint
of hardware we consider C* supported on" and targeting ASF CI to validate that.
I believe the noisy env + low resources on ASF CI currently are lower than
whatever floor we'd reasonably agree on.{quote}
I agree more resources would be great. And we're now on a degraded situation
while some workers are offline getting their 'full HDD' issues fixed. Also
setting some HW baseline seems like a good point.
On the other hand I can't stop finding legit bugs on jenkins runs. Being a
contended env, as opposed to circle, many bugs that circle would never show
come up in jenkins. The problem is being able to find those within all the
noise in jenkins. That's why it is so important to get to green and _keep it_
like that. Otherwise the noise hides the legit failures.
> Raise test timeouts
> -------------------
>
> Key: CASSANDRA-17729
> URL: https://issues.apache.org/jira/browse/CASSANDRA-17729
> Project: Cassandra
> Issue Type: Bug
> Components: Test/unit
> Reporter: Berenguer Blasi
> Assignee: Berenguer Blasi
> Priority: Normal
> Fix For: 4.1-beta
>
>
> We have seen for some time now junits timeout frequently on jenkins. This is
> probably down to it being a very loaded env. On circle we don't observe that
> behavior probably bc it's not so loaded.
> The question is whether it is time to raise timeouts as they might be hiding
> legit failures. As en experiment I raised timeouts in a branch and ran
> jenkins against it. What I see is that the last 4.1 run had 14 failures out
> of which 12 were timeouts. Increasing timeouts reveals what looks to be 9
> legit failures where 2 are timeouts that probably need to be investigated.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
