the systemd units corresponding to the osds which have been removed can be deleted
from the ceph-ansible execution (and mistral workflow status) it looks like the migration into containers and upgrade succeeded even though it took longer than the heat timeout, hence the update failed
the openstack overcloud ceph-upgrade run should be re-attempted passing
the timeout can be increased using, for example
to set a 6hrs timeout, this will bring back the heat stack in good state after which upgrade can be finalized
You should also ensure you're not running into the following issues by updating to the latest patchesets.
Did ensuring the patches from the KCSs in comment #3 and then re-running with --timeout 360 address the issue? If I don't hear back from you in a week I'm close the bug.
The information requested a week ago was not provided so I'm closing this bug as mentioned a week ago. If the information becomes available then feel free to reopen.