Bug#1062356: adios2: flaky autopkgtest (host dependent): times out on big host
Source: adios2 Followup-For: Bug #1062356 The flakey test is adios2-mpi-examples. debian/tests is building and running them manually, and running on only 3 processors (mpirun -n 3) So the problem can't be overload of the machine. I'll just skip insituGlobalArraysReaderNxN_mpi. For reference, upstream is making some changes to make it more reliable to run tests against the installed library, https://github.com/ornladios/ADIOS2/pull/3906 also https://github.com/ornladios/ADIOS2/pull/3820 Not certain that that directly makes insituGlobalArraysReaderNxN_mpi more reliable though.
Bug#1062356: adios2: flaky autopkgtest (host dependent): times out on big host
Source: adios2 Followup-For: Bug #1062356 Can't be quite as simple as just the host machine. https://ci.debian.net/data/autopkgtest/unstable/amd64/a/adios2/41403641/log.gz completed in 9 minutes, while https://ci.debian.net/data/autopkgtest/unstable/amd64/a/adios2/41496866/log.gz failed with timeout. But that was ci-worker13 in both cases. Maybe it's a race condition. Might be simplest to just skip insituGlobalArraysReaderNxN_mpi though I can also review how many CPUs are invoked by the test. Usually safer not to run tests on all 64 available cpus, for instance, if there are that many on the machine.
Bug#1062356: adios2: flaky autopkgtest (host dependent): times out on big host
Source: adios2 Version: 2.9.2+dfsg1-8 Severity: serious User: debian...@lists.debian.org Usertags: flaky Dear maintainer(s), I looked at the results of the autopkgtest of your package. I noticed that it regularly fails. The failures seem related to the host that runs the test. ci-worker13 is a beefy machine [1], while the other amd64 workers are much more moderate [2]. The tests that time out after 2:50 hours seem to all run on ci-worker13 (so, lots of CPU's and lots of RAM), while the other runs only take below 10 minutes (and seem to run on one of the other hosts. Because the unstable-to-testing migration software now blocks on regressions in testing, flaky tests, i.e. tests that flip between passing and failing without changes to the list of installed packages, are causing people unrelated to your package to spend time on these tests. Don't hesitate to reach out if you need help and some more information from our infrastructure. Paul [1] https://metal.equinix.com/product/servers/m3-large/ [2] https://aws.amazon.com/ec2/instance-types/m5/ OpenPGP_signature.asc Description: OpenPGP digital signature
