We've deployed a cluster of 4 nodes with Proxmox and Ceph, with both SSD and HDD devices. The cluster is performing as expected and we are happy with the configuration, but want to make the recovery a bit more robust in case of a major failure. We are also finding out that 3 PVE nodes should be more than enough for our use cases, so will need to remove that 4th node from the cluster. The documentation specify this procedure, so we should be good on that front.
The items that we need to change and that we need help on how to accomplish them are the following:
Thanks,
IB
The items that we need to change and that we need help on how to accomplish them are the following:
- The cluster backplane is 1x10gb nic card. Originally deployed the heartbeat, public and ceph private networks on the same IP address. We then managed to split the Ceph private IP into a VLAN of the same NIC, and we might also move it to a second 10Gb nic once we finalize the physical cluster sizing.
2. We want to increase the number of Ceph nodes, to get more distributed capacity in addition to the PVE cluster servers. How shall we go about adding these additional nodes to the Ceph cluster? Note that we want these machines to be out of the PVE cluster itself. Just Ceph nodes.
Thanks,
IB