Description of problem:
In RHCS 5.1, for the disconnected environment, the below code snippet was added through the fix (https://github.com/ceph/ceph/commit/1f7c13a11e3597edb88b50a9f23a352e83fb24a5) due to which if the container image name is not in the format of <registry>/ceph/ceph the "docker.io" prefix gets added automatically.
 
~~~
    bits = digest.split('/')
    if '.' not in bits[0] or len(bits) < 3:  
        digest = 'docker.io/' + digest
    return digest
~~~
 
However, there was an another fix i.e https://github.com/ceph/ceph/commit/9271e721e671025b340b0a9867842bb6d3b531b1 where the above code snippet has been reworked. So can we have timeline when the fix would be available in the downstream and in which version ?
 
 
Version-Release number of selected component (if applicable): RHCS 5.1
 
 
How reproducible:
If the container image passed in the command `ceph orch upgrade <image>` is not in the format <registry>/ceph/ceph then the "docker.io" prefix get automatically added to the target image name
 

Steps to Reproduce:
1. Deploy the RHCS 5.1 cluster in the **disconnected environment
2. Execute the `orch upgrade` command with the following image name format
~~~
bastionceph5.ceph.com:5000/rhceph-5-rhel8, dummy.registry.com/rhceph-5-rhel8, abc.io/ceph
~~~
 
 
Actual results:
The orch upgrade command adds the prefix 'docker.io' automatically to the image name
 
Expected results:
It should not add the prefix 'docker.io' automatically
 
Additional info:
 
There is an upstream tracker for this issue: https://tracker.ceph.com/issues/53594