Change the cluster network in an existing Ceph cluster

[ropelletier]

When I installed Ceph I had not yet created a separate network for the cluster. I have since done this and want to change the cluster network.

I have three nodes, all are configured as monitors, managers, and metadata servers.

I can't seem to find a known recommended procedure. This cluster is not in production yet, so there are no running VMs. However, I really, really don;t want to tear it down and start from scratch. In a perfect world, I would just need to change the config and restart each node. I have also read that all the OSDs need to be restarted.

I greatly appreciate the assistance.



Here is my config (changed a little from actual config for privacy)

[global]
     auth_client_required = cephx
     auth_cluster_required = cephx
     auth_service_required = cephx
     cluster_network = 192.168.100.5/24
     mon_allow_pool_delete = true
     mon_host = 192.168.100.5 192.168.100.7 192.168.100.8
     ms_bind_ipv4 = true
     ms_bind_ipv6 = false
     osd_pool_default_min_size = 2
     osd_pool_default_size = 3
     public_network = 192.168.100.5/24


[client]
     keyring = /etc/pve/priv/$cluster.$name.keyring

[mds]
     keyring = /var/lib/ceph/mds/ceph-$id/keyring

[mds.district-pve1]
     host = district-pve1
     mds_standby_for_name = pve

[mds.district-pve3]
     host = district-pve3
     mds standby for name = pve


[mds.district-pve4]
     host = district-pve4
     mds_standby_for_name = pve


[mon.district-pve1]
     public_addr = 192.168.100.5


[mon.district-pve3]
     public_addr = 192.168.100.7


[mon.district-pve4]
     public_addr = 192.168.100.8

 

[ropelletier]

Bump. Trying one more time.

 

[thomas.weingart]

Same issue with me. Have you solved it yet?
Thanks in advance.