Seems tests fail on modern distributions (or they run extremely slow)

[BuildStream-Migration-Bot]

See original issue on GitLab
In GitLab by [Gitlab user @jjardon] on Apr 4, 2020, 22:25

• Archlinux: https://gitlab.com/BuildStream/buildstream-docker-images/-/jobs/498943572
• Fedora 32: https://gitlab.com/BuildStream/buildstream-docker-images/-/jobs/498776727
• Ubuntu 20.04: https://gitlab.com/BuildStream/buildstream-docker-images/-/jobs/498827818

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @jjardon] on Apr 5, 2020, 10:20

changed the description

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @jjardon] on Apr 5, 2020, 12:29

People in IRC suggested this can be related with #1173

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @jjardon] on Apr 5, 2020, 12:29

marked this issue as related to #1173

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @jjardon] on Apr 5, 2020, 12:29

removed the relation with #1173

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @jjardon] on Apr 5, 2020, 12:29

marked this issue as related to #1173

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @cs-shadow] on Apr 22, 2020, 00:08

Coverage may be a part of it but it's not the only offender here. I ran this pipeline on Fedora 32 without coverage enabled - https://gitlab.com/BuildStream/buildstream-docker-images/-/jobs/520996831. It hasn't finished yet but I can already see a good amount of tests are failing. Although it has managed to run many more tests than the ones where coverage was enabled (~2% with coverage in 2 hours vs. ~20% in 90 mins without it).

I tried to reproduce this locally and it seems like buildbox-casd is failing to start up in time. Although I am not sure why that would be the case. We almost always start with an empty cache (at least in the pipelines mentioned here) so it should really not take long at all.

When trying manually, I noticed it seemed to be stuck without creating the socket. I gave up after a minute or so.

As I said, I don't know why this is failing though.

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @juergbi] on Apr 22, 2020, 06:36

I tried to reproduce this locally and it seems like buildbox-casd is failing to start up in time. Although I am not sure why that would be the case. We almost always start with an empty cache (at least in the pipelines mentioned here) so it should really not take long at all.

When trying manually, I noticed it seemed to be stuck without creating the socket. I gave up after a minute or so.

This appears to be a gRPC issue: grpc/grpc#21213 grpc/grpc#21280. It is supposedly fixed in gRPC upstream in the v1.26.x, v1.27.x and master branches: grpc/grpc#21662

However, Fedora 32 doesn't include the fix yet: https://bugzilla.redhat.com/show_bug.cgi?id=1812743 https://bugzilla.redhat.com/show_bug.cgi?id=1794566#c9

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @BenjaminSchubert] on Aug 3, 2020, 10:43

mentioned in commit BenjaminSchubert/bst-workspace[Gitlab user @71be778d6af80d469fc70679f30572c29a48ea51]

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @juergbi] on Aug 12, 2020, 20:51

Fedora 32 is now in CI: !2025

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @juergbi] on Aug 13, 2020, 05:39

A test with Ubuntu 20.04 has been successful as well now https://gitlab.com/BuildStream/buildstream-docker-images/-/jobs/683494909

I consider this issue resolved, closing.

[BuildStream-Migration-Bot]

In GitLab by [Gitlab user @tristanvb] on Nov 6, 2020, 09:56

removed the relation with #1173