[ceph-users] Monitor dissapears/stopped after testing monitor-host loss and recovery

Hi, I'm currently testing some disaster scenarios. When removing one osd/monitor host, I see that a new quorum is build without the missing host. The missing host is listed in the dashboard under Not In Quorum, so probably everything as expected. After restarting the host, I see that the osd's come back online and everything appears to be working, however the quorum is still only with two monitors. Looking at the services, I can see that it is somehow stopped. Is this expected and I must start it manually somehow, or should it work? The whole cluster is deployed using cephadm (the node was the initial bootstrap one, if that is important) Greetings, Kai