Let me start by saying, I know a 2-node cluster is not recommended but I only have limited amount of money for my homelab
I also know this isn't a forum for Ceph Support but any help/tips would be greatly appreciated as I'm trying to learn this stuff by myself.
I don't mind having downtime for this procedure.
So I have tried to change the
However, as soon as I do this change, the ceph cluster stops responding and times out. I can't query it with commands like
If I tail /var/log/syslog I get:
I assume this happens because the ceph cluster no longer has quorum. If I revert the changes, everything starts working again.
What am I doing wrong? Or what is the correct procedure of achieving this?
I also know this isn't a forum for Ceph Support but any help/tips would be greatly appreciated as I'm trying to learn this stuff by myself.
I don't mind having downtime for this procedure.
ceph.conf
:
Code:
[global]
auth_client_required = cephx
auth_cluster_required = cephx
auth_service_required = cephx
cluster_network = 172.18.21.200/24
fsid = 083f7640-cb8e-467e-8fa1-4d29291c17aa
mon_allow_pool_delete = true
mon_host = 172.18.21.200 172.18.21.100
osd_pool_default_min_size = 2
osd_pool_default_size = 2
public_network = 172.18.21.200/24
[client]
keyring = /etc/pve/priv/$cluster.$name.keyring
[mds]
keyring = /var/lib/ceph/mds/ceph-$id/keyring
[mds.legion]
host = legion
mds standby for name = pve
[mds.magneto]
host = magneto
mds_standby_for_name = pve
[mon.legion]
public_addr = 172.18.21.100
[mon.magneto]
public_addr = 172.18.21.200
So I have tried to change the
[global].public_network
subnet to 172.18.10.200/24, I changed the [global].mon_host
to the same subnet and I changed the [mon.x].public_addr
to the new subnet. I made sure ceph.conf
was replicated on the 2 hosts.However, as soon as I do this change, the ceph cluster stops responding and times out. I can't query it with commands like
ceph -s
because they just hang.If I tail /var/log/syslog I get:
May 14 10:23:25 legion ceph-mon[9507]: 2020-05-14 10:23:25.109 7f0a877d0700 -1 mon.legion@1(electing) e2 failed to get devid for : fallback method has serial ''but no model
I assume this happens because the ceph cluster no longer has quorum. If I revert the changes, everything starts working again.
What am I doing wrong? Or what is the correct procedure of achieving this?