Bug#1102062: zookeeper: FTBFS: expected: <1> but was: <0>

[tony mancill]

On Thu, Jun 19, 2025 at 06:07:01PM +0200, Santiago Vila wrote:
> severity 1102062 serious
> thanks
> 
> Hi. This package FTBFS around 50% of the time here.
> 
> It may not fail for everybody, but we promised our users that
> they can rebuild the packages from source. A failure rate
> of 50% exceeds the common thresholds used by the RT.
> 
> [ I can still offer a test VM to reproduce if required, but if I
> was the maintainer, I would just disable the flaky test ].
> 
> ( Note: The stable version also FTBFS randomly )

Hi Santiago,

Thank you for performing these test builds.

I'm not able to coerce the test or the build to fail on 3 different
systems in 10 overall attempts. This makes me suspect something related
to the build environment, possibly either the networking stack or
CPU.  The logs you provide show a fairly simple test that creates a
ZK client timing out (taking over 60 seconds) because it never
registers with JMX.  Are sockets scarce in your build environment?
Cores?

[ERROR] Tests run: 2, Failures: 2, Errors: 0, Skipped: 0, Time elapsed: 122.365 s <<< FAILURE! - in org.apache.zookeeper.test.SaslAuthRequiredMultiClientTest

The tests are getting stuck here during some of your builds:
https://salsa.debian.org/java-team/zookeeper/-/blob/master/zookeeper-server/src/test/java/org/apache/zookeeper/test/JMXEnv.java#L100-129

The test completes quickly in all of my trials:

[INFO] Running org.apache.zookeeper.test.SaslAuthRequiredMultiClientTest
[INFO] Tests run: 2, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 1.205 s - in org.apache.zookeeper.test.SaslAuthRequiredMultiClientTest

Please do provide either a VM or more information about the build
environment where this fails and I will work to get the bottom of it.
As far as I can see, there's nothing flaky about the test code.

If there are specific system requirements for a successful build, we
will document those.  And if we can get patch the test to build
successfully in the environment where it fails, we will do that too.

However, I don't see the utility in increasing the severity of the bug
to RC and thereby excluding this package and its multiple reverse
dependencies from trixie.  I think the bug severity should be important.

Thanks,
tony
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/pkg-java-maintainers/attachments/20250622/02b5c74f/attachment.sig>