On 3/16/20 6:25 PM, Jeff Layton wrote:
On Mon, 2020-03-16 at 22:13 +0000, Sage Weil wrote:
There are several of the failures in this run:
http://pulpito.ceph.com/sage-2020-03-16_13:50:19-rados-wip-sage-testing-2020...
_______________________________________________ Sepia mailing list -- sepia@ceph.io To unsubscribe send an email to sepia-leave@ceph.io
Problem with the repos?
This one:
http://qa-proxy.ceph.com/teuthology/sage-2020-03-16_13:50:19-rados-wip-sage-...
2020-03-16T18:50:40.998 INFO:teuthology.orchestra.run.smithi086:> sudo yum install -y kernel 2020-03-16T18:50:43.440 INFO:teuthology.orchestra.run.smithi086.stdout:CentOS-8 - AppStream 0.0 B/s | 0 B 00:00 2020-03-16T18:50:43.440 INFO:teuthology.orchestra.run.smithi086.stderr:Failed to download metadata for repo 'CentOS-AppStream' 2020-03-16T18:50:43.440 INFO:teuthology.orchestra.run.smithi086.stderr:Error: Failed to download metadata for repo 'CentOS-AppStream' 2020-03-16T18:50:43.493 DEBUG:teuthology.orchestra.run:got remote process result: 1
This is insane. I thought we were just having intermittent connectivity issues to the CentOS's infra but we're not even using it anymore. The yum repo conf is using a local mirrorlist [1] and the closest mirror is a couple racks over in the CentOS lab. Then I came across this [2] and implemented this [3]. So there is absolutely no leftover metadata and yum/dnf has 4 different mirrors to connect to. I've gone beyond what I think is reasonable to workaround this so I've filed https://bugzilla.redhat.com/show_bug.cgi?id=1814337 [1] https://github.com/ceph/ceph-cm-ansible/pull/542/commits/85b30c1ca6744839723... [2] https://stackoverflow.com/questions/59993633/yum-dnf-error-failed-to-downloa... [3] https://github.com/ceph/ceph-cm-ansible/pull/549/commits/4de5c5b4d3046e0b193...