Full Mesh for Ceph Storage

[ScottDavis]

Just completed following this: https://pve.proxmox.com/wiki/Full_Mesh_Network_for_Ceph_Server.

Confirmed new openfabric network is configured on the correct interfaces. Monitors are showing as up on each node, but I am getting 'got timeout (500) when viewing ceph under cluster?

I don't want to start building OSD / Pools till I've fixed this error which I'm assuming is down to cluster conf.?

*edit* fixed this... now on to the next issue :/

 

[ScottDavis]

Restarting the primary node seems to have sorted it. However, I can see the node manager is not happy.

I destroyed the manager after adding the other two nodes under manager successfully, however when I go to add the node manager I destroyed I get status "unknown" and it times out when adding.

Could it be due to some clock skew detected?

Out of curiousity created OSD on a node ... replicates to two nodes, but not the primary node that is complaining.