2024-08-21T01:11:44.860Z | <Samuel Just> 5843c6b04bacf9a7c981fca5d874ab3400f855db introduced the nvmeof/gateway submodule:
```+[submodule "src/nvmeof/gateway"]
+ path = src/nvmeof/gateway
+ url = <https://github.com/ceph/ceph-nvmeof.git>
+ fetchRecurseSubmodules = false
+ shallow = true``` |
2024-08-21T01:12:20.423Z | <Samuel Just> 5843c6b04bacf9a7c981fca5d874ab3400f855db introduced the nvmeof/gateway submodule:
```+[submodule "src/nvmeof/gateway"]
+ path = src/nvmeof/gateway
+ url = <https://github.com/ceph/ceph-nvmeof.git>
+ fetchRecurseSubmodules = false
+ shallow = true```
I suspect the shallow = true is the problem -- it seems to prevent git submodule update from actually finding the right sha1 |
2024-08-21T01:16:25.081Z | <Samuel Just> <https://tracker.ceph.com/issues/67640> |
2024-08-21T01:27:29.942Z | <Samuel Just> <https://github.com/ceph/ceph/pull/59366> |
2024-08-21T07:55:36.405Z | <Jan Horacek> hi @Casey Bodley rook issue that looks like the same issue is here <https://github.com/rook/rook/issues/14434> |
2024-08-21T07:56:12.572Z | <Jan Horacek> i will report it to ceph tracker |
2024-08-21T14:07:55.152Z | <nehaojha> The Ceph User Dev Monthly meeting is happening now <https://meet.jit.si/ceph-user-dev-monthly>. |
2024-08-21T15:26:25.038Z | <Casey Bodley> weekly rgw meeting soon in [ <https://pad.ceph.com/p/rgw-weekly](https://meet.google.com/mmj-uzzv-qce> (new meeting link) ) |
2024-08-21T17:59:56.656Z | <Matthews Jose> Hi everyone,
I’ve been exploring the Ceph codebase while trying to understand how RADOS sends a read request to the OSD. I have a few specific questions:
1. I noticed that the `CEPH_MSG_OSD_OP` constant is defined in `ceph_fs.h`. This is a bit surprising since I expected this to be in an OSD-specific header, such as one more closely associated with OSD operations. Could someone clarify why it is defined in this file instead?
2. Additionally, I want to confirm how the message type field is populated in the Ceph header when a read operation is performed. From what I understand, for a read request, the message type in the Ceph header is set to `CEPH_MSG_OSD_OP` (value 42). Is this correct, and could anyone provide more context on how this is handled during an OSD operation?
Any insight into these points would be greatly appreciated!
Thanks in advance! |
2024-08-21T18:45:08.151Z | <Casey Bodley> for 1, i'm guessing that's for historical reasons. my understanding is that ceph_fs.h is one of the few headers that are shared between the ceph repo and its kernel clients |
2024-08-21T18:49:37.069Z | <Casey Bodley> for 2, you should be able to find most of the low-level client-side details in [class Objecter](https://github.com/ceph/ceph/blob/main/src/osdc/Objecter.h) which is used by librados |
2024-08-21T18:50:42.522Z | <Casey Bodley> there's a lot going on inside the OSD to handle requests, but [PrimaryLogPG::do_osd_ops()](https://github.com/ceph/ceph/blob/main/src/osd/PrimaryLogPG.cc#L5989) is probably a good place to start? |
2024-08-21T18:54:08.523Z | <Matthews Jose> Thank you for your reply and the references. |
2024-08-21T20:12:23.267Z | <Joseph Mundackal> is the owner/created/author of the PR of a tracker (eg: <https://tracker.ceph.com/issues/65791>) expected to have permissions to change the status of a tracker to "pending backport" ? |
2024-08-21T20:56:02.127Z | <Joseph Mundackal> thanks @nehaojha
do you know what controls the permissions for this? and is that meant to be locked down? |
2024-08-21T20:59:26.121Z | <nehaojha> I think they might need a dev role in the tracker or something |
2024-08-21T21:00:26.285Z | <Joseph Mundackal> any docs on how to get that setup? |
2024-08-21T22:31:20.484Z | <nehaojha> @Adam Kraitman do you know there's any documentation around it or is it just redmine documentation? |