Hi Paride,
After the mass-rebuild & retag needed to address
https://bugs.launchpad.net/bugs/1930094, I tried triggering a few of our
tests in the CI. Some more tests are passing now, and we're on our way
to achive 100% success rate soon. There are few things that need to be
adjusted in the setup, so I decided to write this message to let you
know:
1) The apache2 tests are not passing. Initially I thought this was
because of the bug mentioned above, but it's actually something else: we
will need to explicitly enable IPv6 support in the docker client in
order to run one of the unit tests. I had the same problem locally, and
here's what I did:
# cat /etc/docker/daemon.json
{
"ipv6": true,
"fixed-cidr-v6": "fd00::/80"
}
# systemctl restart docker.service
I tried (not too hard) to enable IPv6 when creating the docker network
that is used to run the tests (via the --ivp6 option), but it didn't
work OOTB.
The nginx test suffers from the same problem, *but* strangely enough I
wasn't able to make its IPv6 test pass even with the modifications
above.
2) The s390x machine/VM/container needs an update, I think. It's still
affected by that s390x kernel problem, as can be seen here:
https://jenkins.ubuntu.com/server/view/oci/job/oci-unit-memcached/DOCKER_NAMESPACE=ubuntu,DOCKER_REGISTRY=public.ecr.aws,DOCKER_TAG=edge,nodes=metal-s390x/lastBuild/console
3) It seems some of the test definitions need to be updated because of
the s/master/main/ change. I haven't checked them all, but I noticed
that at least the grafana and redis tests are failing because of this.
I think this is everything.
Thanks a lot,