Connection refused from 2 of 4 nodes on a cluster

[filippoclikkami]

Hi, i get the error message "595 Connection refused" if i try to manage 2 of a 4 nodes cluster. This is a production cluster and every node comes with 1 dedacted 10gig nic which one with 2 ports, one for HA and one for Ceph. The manage network is on default 1gig nic, checking on logs i'll see that barely often RADOS gives offline alert from these 2 nodes but they still work on the VM side. I try to change switchs but doesn't solve anything, any suggestions? If i change the manage nic have i to re-add the nodes on the cluster or it works autamtically?


PVE 7.4-17
Ceph 17.2.7

 

[michel.seicon]

Hello.
You can adjust your cluster to be able to manage it

pvecm expected 2

Yes, after the nodes communicate, they synchronize, just upload the networks to do so.

Climb one at a time.