Bug #70834
openmgr/vol: add "snapshot getpath" cmd for subvol v1
0%
Description
#70815 is about adding "ceph fs subvolume snapshot getpath" command to get the
path of a subvolume snapshot for subvolume v2. Under this ticket, this same
needs to be done but for subvolume v1 instead. This change must follow changes
for ticket #70833 since it won't be possible to test the code for this ticket
unless v1 is tested in fs:volumes.
Updated by Rishabh Dave 12 months ago
- Related to Enhancement #70815: mgr/vol: add a command to get the path of snapshots added
Updated by Rishabh Dave 12 months ago
- Related to Bug #70833: mgr/vol: add support for legacy, and v1 subvol version in fs:volumes added
Updated by Venky Shankar 11 months ago
OK, so thinking on this again: Getting snapshot path for v1 subvolumes isn't implemented, which I understand is kind of dependent on testing v1 subvolumes which could take some time. So, I propose that we do what's required for the command to work for v1 subvolume and introduce basic test covering that.
https://tracker.ceph.com/issues/70833 can be worked on post that which would add full testing with all subvolume versions (including v1 which not always gets upgraded to v2).
Updated by Rishabh Dave 11 months ago · Edited
Venky Shankar wrote in #note-3:
OK, so thinking on this again: Getting snapshot path for v1 subvolumes isn't implemented, which I understand is kind of dependent on testing v1 subvolumes which could take some time. So, I propose that we do what's required for the command to work for v1 subvolume and introduce basic test covering that.
https://tracker.ceph.com/issues/70833 can be worked on post that which would add full testing with all subvolume versions (including v1 which not always gets upgraded to v2).
vshankar So essentially what we want in the test is that a v1 subvol is created and it's not upgraded at any point during the test. And then on this un-upgraded v1 subvol, we make sure @snapshot getpath works fine. Correct?
Updated by Rishabh Dave 11 months ago
Rishabh Dave wrote in #note-4:
Venky Shankar wrote in #note-3:
OK, so thinking on this again: Getting snapshot path for v1 subvolumes isn't implemented, which I understand is kind of dependent on testing v1 subvolumes which could take some time. So, I propose that we do what's required for the command to work for v1 subvolume and introduce basic test covering that.
https://tracker.ceph.com/issues/70833 can be worked on post that which would add full testing with all subvolume versions (including v1 which not always gets upgraded to v2).
vshankar So essentially what we want in the test is that a v1 subvol is created and it's not upgraded at any point during the test. And then on this un-upgraded v1 subvol, we make sure @snapshot getpathworks fine. Correct?
Had a chat with Venky. Path ahead is to have a test (and of course having supporting source code for it) that constructs a v1 subvol and then checks that ENOSTSUPP is returned by snapshot getpath command against it (it = v1 subvol).
Updated by Rishabh Dave 11 months ago
- Status changed from New to Fix Under Review
- Pull request ID set to 63123
Updated by Rishabh Dave 11 months ago · Edited
- Pull request ID changed from 63123 to 63127
Rishabh Dave wrote in #note-5:
Rishabh Dave wrote in #note-4:
Venky Shankar wrote in #note-3:
OK, so thinking on this again: Getting snapshot path for v1 subvolumes isn't implemented, which I understand is kind of dependent on testing v1 subvolumes which could take some time. So, I propose that we do what's required for the command to work for v1 subvolume and introduce basic test covering that.
https://tracker.ceph.com/issues/70833 can be worked on post that which would add full testing with all subvolume versions (including v1 which not always gets upgraded to v2).
vshankar So essentially what we want in the test is that a v1 subvol is created and it's not upgraded at any point during the test. And then on this un-upgraded v1 subvol, we make sure @snapshot getpathworks fine. Correct?Had a chat with Venky. Path ahead is to have a test (and of course having supporting source code for it) that constructs a v1 subvol and then checks that ENOSTSUPP is returned by
snapshot getpathcommand against it (it = v1 subvol).
The work for this ticket has taken a different turn. We are now supportingsnapshot getpath for v1 and legacy subvolumes as well. This is possible with
current source code, all we need to do is to add tests for this command against
v1 and legacy subvolumes. This is what the new PR does.
Therefore, https://github.com/ceph/ceph/pull/63123 has been closed in favour of
https://github.com/ceph/ceph/pull/63127.
Updated by Rishabh Dave 10 months ago
- Status changed from Fix Under Review to Pending Backport
Updated by Rishabh Dave 10 months ago
- Copied to Backport #71276: squid: mgr/vol: add "snapshot getpath" cmd for subvol v1 added
Updated by Rishabh Dave 10 months ago
- Copied to Backport #71277: tentacle: mgr/vol: add "snapshot getpath" cmd for subvol v1 added
Updated by Rishabh Dave 10 months ago
- Copied to Backport #71278: reef: mgr/vol: add "snapshot getpath" cmd for subvol v1 added
Updated by Rishabh Dave 10 months ago
- Tags (freeform) set to backport_processed
Updated by Upkeep Bot 9 months ago
- Merge Commit set to beed49e5e5a3fa6bb3cd0ff3a476a495b7810424
- Fixed In set to v20.3.0-289-gbeed49e5e5a
- Upkeep Timestamp set to 2025-07-08T14:46:51+00:00
Updated by Upkeep Bot 8 months ago
- Fixed In changed from v20.3.0-289-gbeed49e5e5a to v20.3.0-289-gbeed49e5e5a3
- Upkeep Timestamp changed from 2025-07-08T14:46:51+00:00 to 2025-07-14T15:21:01+00:00
Updated by Upkeep Bot 8 months ago
- Fixed In changed from v20.3.0-289-gbeed49e5e5a3 to v20.3.0-289-gbeed49e5e5
- Upkeep Timestamp changed from 2025-07-14T15:21:01+00:00 to 2025-07-14T20:45:35+00:00