2024-06-17T07:59:53.610Z

<Zac Dover> @Guillaume Abrioux, will I take anyone by unwelcome surprise by merging <https://github.com/ceph/ceph/pull/58004> ?

2024-06-17T08:06:12.050Z

<Guillaume Abrioux> @Zac Dover please wait

2024-06-17T08:06:27.007Z

<Guillaume Abrioux> I prefer ceph-volume maintainers to merge ceph-volume related PRs

2024-06-17T08:50:45.649Z

<Zac Dover> @Guillaume Abrioux: I will wait. Thank you for the speedy reply.

2024-06-17T08:52:10.640Z

<Zac Dover> <https://github.com/ceph/ceph/pull/58005>
<https://github.com/ceph/ceph/pull/58006>

Are also outstanding, and are part of the set of backports necessary to get the readthedocs builds back up and running.

2024-06-17T12:35:13.811Z

<Sebastian Wagner> can we take care of the spamming?

2024-06-17T12:35:21.917Z

<Sebastian Wagner> can we take care of the spam?

2024-06-17T13:33:00.137Z

<Stefan Kooman> @Zac Dover Can you please check this tracker issue: <https://tracker.ceph.com/issues/56650>. And use the suggested "workaround"  in the documentation: <https://docs.ceph.com/en/latest/rados/operations/stretch-mode/>? Not sure when the fix will get merged upstream. So in the mean time every operator that sets up a "stretch mode" cluster following said documentation will hit this issue.

2024-06-17T13:50:38.645Z

<Stefan Kooman> @Laura Flores <https://tracker.ceph.com/issues/51729#note-26> <- This note suggests that there are changes to "verify_upmap" code to handle issues #51729 issue and #56650. The reason for pinging you is that I have found out it's possible to create "valid" upmaps that break the CRUSH rule policy when using the "workaround" CRUSH rule (stretch_replication_rule) suggestion in #56650 (see also <https://github.com/TheJJ/ceph-balancer/issues/41#issuecomment-2173349109>). So I would like to verify that the new "verify_upmap" code also detects these "valid" upmaps and make it impossible to set these upmaps and break CRUSH policy.

2024-06-17T13:59:55.804Z

<Zac Dover> @Stefan Kooman, I'll generate a PR before noon tomorrow New York time, and I'll designate you and Radek as the reviewers of it.