To Reproduce Steps to reproduce the behaviour:

1. In a RHOSO environnement:

Cu created the following ceph keyring :

cephadm shell – \
   ceph auth add client.openstack \
     mgr 'allow *' \
        mon 'allow r' \
        osd 'allow class-read object_prefix rbd_children, allow rwx pool=vms, allow rwx pool=volumes, allow rwx pool=images'

Like the one mentioned in official docs 3.1.4

2. But rbd lock is not released when the compute crashes, so it seems some caps are missing.

3. The update the keyring with profile rbd like mentioned in another chapter of official documentation 3.9.7 :

bash-5.1$ cat /etc/ceph/ceph.client.openstack.keyring
[client.openstack]
   key = "<redacted>"
   caps mgr = allow *
   caps mon = profile rbd
   caps osd = profile rbd pool=vms, profile rbd pool=volumes, profile rbd pool=backups, profile rbd pool=images

And now they don't face any issues.

Expected behavior

• I expect to see same definition on both paragraph

Device Info (please complete the following information):

• RHOSO18.0.2
• RHCeph7.1.1

Bug impact

• This customer don't have now the issue but want to be sure rbd profile is the right one.
• But other customer can be impacted by this we should update the docs asap.

Known workaround

• The update the keyring with profile rbd like mentioned in another chapter of official documentation 3.9.7.

Additional context

• Also, ceph docs advice to use rbd profile.