Bug 2314146

.Sometimes ceph-mgr modules are temporarily unavailable and their commands fail Occasionally, the balancer module takes a long time to load after a ceph-mgr restart. As a result, other ceph-mgr modules can become temporarily unavailable and their commands fail. For example, ---- [ceph: root@host01 /]# ceph crash ls Error ENOTSUP: Warning: due to ceph-mgr restart, some PG states may not be up to date Module 'crash' is not enabled/loaded (required by command 'crash ls'): use `ceph mgr module enable crash` to enable it ---- As a workaround, in cases after a ceph-mgr restart, commands from certain ceph-mgr modules fail, check the status of the balancer, using the ceph balancer status command. This might occur, for example, during an upgrade. If the balancer was previously active `"active": true` but now it is marked as `“active": false`, check its status until it is active again, then rerun the other ceph-mgr module commands. In other cases, try to turn off the balancer ceph-mgr module. ---- ceph balancer off ---- After turning off the balancer, rerun the other ceph-mgr module commands.