2024-10-29T15:21:03.746Z

<gregsfortytwo> so we have a downstream bz that using cephfs-mirror to copy a subvolume setup doesn’t produce subvolumes on the other side: <https://bugzilla.redhat.com/show_bug.cgi?id=2267050>.
But the tracker ticket @Milind Changire created for this <https://tracker.ceph.com/issues/65157> got closed by @jcollin as “Can’t Reproduce”. Can you guys discuss this so we can either close out the bz or else have enough of a task to schedule it?

2024-10-29T15:25:31.200Z

<Milind Changire> sure, will take a look

2024-10-29T16:20:10.194Z

<Venky Shankar> As I remember, the effort was put in the back burner since it wasn't a prio, but, yeh, can't reproduce is not a correct resolution to the tracker.

2024-10-29T16:25:09.205Z

<Igor Golikov> Hi I see something very very weird.. I am trying to understand the reason for multiple test failures in my teuthology run. Specifically, I have one test, that fails with message that should not be present in the current code.
[https://github.com/ceph/ceph/commit/9962772358048a98a6e871dccf1bfd0a15b4d791#diff-d3a5b657546e321cea1925b2abe92[…]c9326f6ea362549f5e6fb3c0edb9dL1311](https://github.com/ceph/ceph/commit/9962772358048a98a6e871dccf1bfd0a15b4d791#diff-d3a5b657546e321cea1925b2abe9238c907c9326f6ea362549f5e6fb3c0edb9dL1311)
This line shows the message `ss << "module '" << module << "' cannot be disabled (always-on)"`  has been changed to `ss << "module '" << module << "' cannot be disabled (always-on), use " << "'ceph mgr module force disable' command to disable an always-on module"`
The return value has been changed from EINVAL to EPERM. This is the reason the test I see fails, because it expects the new return value (EPERM).
Is it possible that some of the nodes in my teuthology run use old version of the code? Instead of the SHA specified by -S ?

2024-10-29T16:25:59.627Z

<Igor Golikov> @gregsfortytwo this is apparently the reason for multiple failures. it seems that some of the jobs run mixed version of code...