This bug was initially created as a copy of Bug #2211324 I am copying this bug because: Description of problem: ----------------------- - After upgrading the cluster from RHCS 4.3z1 (baremetal) to RHCS 5.3z3 / RHCS 5.3z2, if we run cephadm-preflight playbook to install the latest ceph-common and cephadm packages on the ceph nodes, it stops the ceph.target service, which in turn stops all the ceph services running on the host. This happens only when ceph rpms like ceph-common,ceph-base,ceph-mon,ceph-osd etc from older version (RHCS 4.3z1) still exists on the hosts (as cluster is migrated from baremetal to container). Version-Release number of selected component (if applicable): ------------------------------------------------------------- RHCS 5.3 How reproducible: ----------------- Every time. Steps to Reproduce: -------------------- 1. Deploy RHCS 4.3z1 baremetal cluster 2. Convert the Ceph services to containerized 3. Upgrade the cluster to RHCS 5.3z2 / RHCS 5.3z3 4. Run cephadm-preflight playbook to upgrade the ceph-common and cephadm package on the host. Actual results: --------------- The ceph packages is upgraded but all ceph services on the host are stoped. Expected results: ----------------- The ceph packages should be upgraded and no services should be impacted.