Comment 4 for bug 1928690

That works on my working cluster:

(nova-compute)[root@chrnc-dev-os-compute-01 /]# ceph --id nova health
HEALTH_OK

Fails on the upgraded one:

(nova-compute)[nova@chrnc-void-testupgrade-compute-1-replace /]$ ceph --id nova health
[errno 2] error connecting to the cluster