ceph - ceph-devel - 2024-08-05

Timestamp (UTC)Message
2024-08-05T02:30:09.879Z
<Xinliang Liu> Hi <@U04JGS8QR9U>, I tried, but it is still not working: https://files.slack.com/files-pri/T1HG3J90S-F07G4F0NYRW/download/image.png
2024-08-05T02:34:03.539Z
<Xinliang Liu> And the branch pushed by my colleague only builds x86 rpms not aarch64, what can I do if I want to build aarch64 rpms?
<https://3.chacra.ceph.com/r/ceph/ceph-nvmeof-mon-arm64-testin/0adaeecf622b3867ee27e9fabeb939831c4967dc/centos/9/flavors/default/aarch64/>
2024-08-05T06:24:48.834Z
<Rico> Hello, any news about <https://tracker.ceph.com/issues/64213> ? I've been waiting 18.2.4 expecting fixes but it seems there's nothing related in changelog 😞
2024-08-05T11:37:22.838Z
<Rost Khudov> Hello folks, when I am building ceph with cmake command, I can see that it is trying to find Python 2 times
The first time it finds `Found Python3:` which is from main [CMakeLists.txt](https://github.com/ceph/ceph/blob/main/CMakeLists.txt) file
but the second time it finds `Found Python:` (right after `Found Cython`) which can cause a problem if there are multiple Python versions
Why in one case it tries to find Python3 and the second time it tries to find just Python? Shouldn't it be Python3 both times?
2024-08-05T12:17:19.971Z
<Andrea Bolzonella> Hi everyone,
We have a situation in one cluster where objects seem to be located in the wrong placement group (PG). Some objects that should be listed in a specific PG are actually mapped to another PG (confirmed using “ceph osd map”). The cluster is running Mimic and we are talking about millions of objects. Any suggestions on how to fix this?
Thanks!
2024-08-05T13:34:11.968Z
<Adam Kraitman> Hey it looks like you didn't accept the invitation: https://files.slack.com/files-pri/T1HG3J90S-F07FLPSTM36/download/image.png
2024-08-05T13:36:01.659Z
<Rost Khudov> Does anybody encounter following error during ceph build?
```Found Python: /usr/bin/python3 (found version "3.9.7") found components: Interpreter 
CMake Error at /usr/share/cmake/Modules/FindPackageHandleStandardArgs.cmake:230 (message):
  Could NOT find Protobuf (missing: Protobuf_LIBRARIES Protobuf_INCLUDE_DIR)
Call Stack (most recent call first):
  /usr/share/cmake/Modules/FindPackageHandleStandardArgs.cmake:594 (_FPHSA_FAILURE_MESSAGE)
  /usr/share/cmake/Modules/FindProtobuf.cmake:646 (FIND_PACKAGE_HANDLE_STANDARD_ARGS)
  src/CMakeLists.txt:901 (find_package)```
2024-08-05T13:45:28.412Z
<Joseph Mundackal> missing portobug libraries?
2024-08-05T13:45:35.038Z
<Joseph Mundackal> missing portobuf libraries?
2024-08-05T13:46:22.210Z
<Rost Khudov> looks like yes, trying to understand when it was added
I suspect this PR <https://github.com/ceph/ceph/pull/54671>
checking
2024-08-05T13:58:14.860Z
<yuriw> reef builds red for long time <https://ceph-storage.slack.com/archives/C1HFJ4VTN/p1722862310211579?thread_ts=1722305916.680979&cid=C1HFJ4VTN>
2024-08-05T14:25:24.800Z
<Guillaume Abrioux> @Zac Dover hello
2024-08-05T14:25:30.697Z
<Guillaume Abrioux> any update?
2024-08-05T15:01:59.970Z
<Zac Dover> no, which branch are you trying to rebase on quincy?
2024-08-05T15:02:49.056Z
<Guillaume Abrioux> <https://github.com/ceph/ceph/pull/58221/files> this one
2024-08-05T15:03:08.983Z
<Guillaume Abrioux> if you checkout my branch, and take a look at `doc/radosgw/adminops.rst`
2024-08-05T15:03:28.977Z
<Guillaume Abrioux> you will see it is different
2024-08-05T16:34:46.723Z
<Zac Dover> I'm looking into it now.
2024-08-05T16:37:49.011Z
<Zac Dover> `zdover@NUC8i7BEH:~/Documents/ceph$ git rebase ceph/quincy`
`Successfully rebased and updated refs/heads/wip-66025-quincy.`
2024-08-05T16:37:53.504Z
<Guillaume Abrioux> yeh
2024-08-05T16:37:54.988Z
<Guillaume Abrioux> i just tried
2024-08-05T16:37:57.611Z
<Zac Dover> I can force-push if you like.
2024-08-05T16:37:58.658Z
<Guillaume Abrioux> i must have done something bad
2024-08-05T16:38:05.701Z
<Guillaume Abrioux> because i can rebase it too
2024-08-05T16:38:09.596Z
<Zac Dover> I have no trouble with it in my local working copy.
2024-08-05T16:38:31.645Z
<Zac Dover> I mean that I have checked out your branch and then rebased on quincy, and I had no trouble.
2024-08-05T16:38:34.860Z
<Guillaume Abrioux> ```hint: Resolve all conflicts manually, mark them as resolved with
hint: "git add/rm <conflicted_files>", then run "git rebase --continue".
hint: You can instead skip this commit: run "git rebase --skip".
hint: To abort and get back to the state before "git rebase", run "git rebase --abort".
Could not apply b8bbecb7a7a... doc/radosgw/adminops: add latest go ceph library to adminops doc
➜  ceph git:(85224d90e0d) ✗ git status
interactive rebase in progress; onto 85224d90e0d
Last command done (1 command done):
   pick b8bbecb7a7a doc/radosgw/adminops: add latest go ceph library to adminops doc
Next commands to do (5092 remaining commands):
   pick 420071be6b7 crimson/os/seastore/journal: convert to seastore logging
   pick 81f6e4e82e7 crimson/os/seastore: convert ExtentReader to seastore logging
  (use "git rebase --edit-todo" to view and edit)
You are currently rebasing branch 'wip-66025-quincy' on '85224d90e0d'.
  (fix conflicts and then run "git rebase --continue")
  (use "git rebase --skip" to skip this patch)
  (use "git rebase --abort" to check out the original branch)

Unmerged paths:
  (use "git restore --staged <file>..." to unstage)
  (use "git add <file>..." to mark resolution)
        both modified:   doc/radosgw/adminops.rst```
2024-08-05T16:38:45.454Z
<Guillaume Abrioux> i've just re-tested and it passes now...
2024-08-05T16:39:05.660Z
<Guillaume Abrioux> i must have done something bad when i tried last week i suppose
2024-08-05T16:39:25.326Z
<Guillaume Abrioux> ```$ git pull --rebase upstream quincy
Warning: Permanently added the RSA host key for IP address '140.82.114.3' to the list of known hosts.
From [github.com:ceph/ceph](http://github.com:ceph/ceph)
 * branch                    quincy     -> FETCH_HEAD
First, rewinding head to replay your work on top of it...
Applying: ceph-volume: create LVs when using partitions```
2024-08-05T16:39:27.574Z
<Guillaume Abrioux> 🤷‍♂️
2024-08-05T16:39:31.508Z
<Zac Dover> There was a week in 2021 when my local working copy was completely incoherent, and I still don't know why.
2024-08-05T16:40:24.057Z
<Zac Dover> For what it's worth, I never once during this saw any conflict in `doc/radosgw/adminops.rst`.
2024-08-05T16:41:56.146Z
<Zac Dover> ¯\_(ツ)_/¯
2024-08-05T16:42:09.483Z
<Guillaume Abrioux> yeh, not sure what i've done last friday
2024-08-05T16:42:21.631Z
<Guillaume Abrioux> but today i can rebase it
2024-08-05T16:42:29.854Z
<Guillaume Abrioux> sorry for the noise 🤷‍♂️
2024-08-05T18:56:59.192Z
<Rolando> I’ve got a similar issue, can somebody gives advice about the recommended value for `mon_osd_min_down_reporters`  in an isolated networking problem?
2024-08-05T20:12:30.055Z
<Matthews Jose> Hi everyone,
In Ceph, how should a replica OSD handle read requests when there is a possibility that it might have missed REPOP messages from the primary OSD due to network latency or failures, especially when there are no clear indications of pending write acknowledgments or missed messages? Specifically, what mechanisms can a replica OSD use to ensure it does not serve stale data, and should it always forward read requests to the primary OSD under such uncertain conditions?
2024-08-05T21:09:43.872Z
<alvaro.soto> where are the reporters located? what subtree?
2024-08-05T21:11:16.592Z
<alvaro.soto> by {number_of_host} you mean number of hosts in the cluster?
2024-08-05T21:12:40.490Z
<alvaro.soto> off the top of my head, I'll say, on a large cluster it will take forever to detect a real osd down
2024-08-05T21:14:02.247Z
<alvaro.soto> so => data corruption and/or data loss
2024-08-05T21:15:19.883Z
<alvaro.soto> recommended value is 2, but I'l also say it depends on the replication value, or ec configuration
2024-08-05T22:34:28.152Z
<Rolando> subtree is at host level
2024-08-05T22:35:03.982Z
<Rolando> {number_of_host} is the total number of hosts in the cluster
2024-08-05T22:35:37.617Z
<Rolando> some pools have replica 3, some have EC 4+2
2024-08-05T22:35:51.266Z
<Rolando> total hosts in cluster = 8

Any issue? please create an issue here and use the infra label.