Bug 2349154

Description of problem:

With some CU we have to help clean up orphaned subvolumes, oftentimes this happens when OCP has cloning issues and we create a lot of pendings which get eventually created but then never consumed by OCP (current Customer case, around 100TB of data in 3700 unused subvolumes, most of them clones. With RBD we can follow a chain of volume->snap->clone->snap->clone etc. but with CephFS this data is not available. 

During cloning we can see source subvolume and snapshot data, but those get lost as soon as the clone has finished:

Test:
```
[quickcluster@mgmt-0 ceph-odf]$ ./cephfs-volume-tool --list-clones

 *********************************************************************************************
  Please note that this is not to be treated as a Red Hat official binary / software,         
  and feel free to go through the source code to look for what's happening behind the scenes. 
 *********************************************************************************************

./cephfs-volume-tool - Version: 4

For details please see the KCS https://access.redhat.com/solutions/7080701

[100%] 3 of 3 subvolumes. Found: 'P': 0 'In-P': 1 'C': 0. Action taken: 0 'P' canceled. 0 'In-P' canceled. 0 'C' Removed.                                                                                                                                                                                                                                                                                                                                                                                  
Log file:
cat ./cephfs-volume-tool_2025-03-01T08:25:29-05:00.log | column -t -s $'\t'

Cmd file:
cat ./cephfs-volume-tool_2025-03-01T08:25:29-05:00.cmd

Finished.
[quickcluster@mgmt-0 ceph-odf]$ cat ./cephfs-volume-tool_2025-03-01T08:25:29-05:00.log | column -t -s $'\t'
2025-03-01T08:25:30-05:00  not-canceling  test-delete                  95  'clone status' is not supported on volume type 'subvolume'.
2025-03-01T08:25:30-05:00  not-canceling  test-clone-from-test-delete  0   complete
2025-03-01T08:25:31-05:00  not-canceling  test-clone2-from-clone       0   in-progress
```

We see cloning is in progress.


Looking at the new .meta file:
```
[quickcluster@mgmt-0 ceph-odf]$ sudo cat   /mnt/cephfs/volumes/csi/test-clone2-from-clone/.meta
[GLOBAL]
version = 2
type = clone
path = /volumes/csi/test-clone2-from-clone/55e13b9a-780b-4534-8b1c-16ed122fb5d5
state = in-progress

[source]
volume = cephfs
group = csi
subvolume = test-clone-from-test-delete
snapshot = test-clone-from-test-delete@snapshot
```

We see all the relevant data in `[source]`.



In addition to that I found this strange behaviour during the cloning process:
```
[quickcluster@mgmt-0 ceph-odf]$ sudo ceph fs clone status --clone_name test-clone2-from-clone --vol_name cephfs --group_name csi 
{
  "status": {
    "state": "in-progress",
    "source": {
      "volume": "cephfs",
      "subvolume": "test-clone-from-test-delete",
      "snapshot": "test-clone-from-test-delete@snapshot",
      "group": "csi"
    },
    "progress_report": {
      "percentage cloned": "92.233%",
      "amount cloned": "1.5G/1.7G",                        <---- 1.5G out of 1.7G, that does look right.
      "files cloned": "60k/57k"                            <---- 60k out of 57k? this does *not* look right.
    }
  }
}
```

After the cloning has finished, checking again the .meta file:
```
[quickcluster@mgmt-0 ceph-odf]$ sudo cat   /mnt/cephfs/volumes/csi/test-clone2-from-clone/.meta
[GLOBAL]
version = 2
type = clone
path = /volumes/csi/test-clone2-from-clone/55e13b9a-780b-4534-8b1c-16ed122fb5d5
state = complete
```

All source information from where the clone was created is gone. 


Engineering Q:

1. Please retain the source data information for the clone in some form in the meta file, this can help establishing from which original subvolume the clone was taken.
2. Check the files cloned output, which seems weird.


Version-Release number of selected component (if applicable):
This cluster has a test image from Guillaume where I am testing a new feature he developed (creating separate DB and BLOCK on the same NVMe device), therefore version is:
ceph version 19.3.0-7350-g7fcd9cba (7fcd9cbac04f022c8a95dfed302e308881bc1da4) squid (dev)


But at least the issue that source information is gone from .meta files after the cloning has completed I have seen since RHCS 5 days.


How reproducible:
Always

Steps to Reproduce:
N/A

Actual results:
N/A

Expected results:
N/A

Additional info: