Description of problem:
following pre-task in the rolling-upgrade.yml playbook fails with NVME devices, where the name of the service is like ceph-osd:
349     - name: get osd unit names - container
350       shell: systemctl list-units | grep -E "loaded * active" | grep -oE "ceph-osd@([0-9]{1,}|[a-z]+).service"

https://github.com/ceph/ceph-ansible/blob/v3.2.8/infrastructure-playbooks/rolling_update.yml#L349
-----------
$ systemctl list-units | grep -E "loaded * active" | grep -oE "ceph-osd@([0-9]{1,}|[a-z]+).service"
$ echo $?
1
-----------


there are two ways to modify it:
A) modified the line to be able to match NVME devices directly:
350       shell: systemctl list-units | grep -E "loaded * active" | grep -oE "ceph-osd@([0-9]{1,}|[a-z]+|nvme.*).service"
$ systemctl list-units | grep -E "loaded * active" | grep -oE "ceph-osd@([0-9]{1,}|[a-z]+|nvme.*).service"
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd

B) Edit the reg exp for named containers after devices:
350       shell: systemctl list-units | grep -E "loaded * active" | grep -oE "ceph-osd@([0-9]{1,}|[a-z0-9]+).service"

"[a-z0-9]+" should work for any combination of characters "a-z" and "0-9" of length 1 character and more, that should include nvmeXn1 too.

$ systemctl list-units | grep -E "loaded * active" | grep -oE "ceph-osd@([0-9]{1,}|[a-z0-9]+).service"
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd
ceph-osd


Version-Release number of selected component (if applicable):
Ceph-ansible-3.28

How reproducible:
always

Steps to Reproduce:
1. deploy containerized ceph with nvme devices
2. run  rolling-upgrade.yml playbook
3. fails for nvme named sevices

Actual results:
fail

Expected results:
not fail

Additional info: