Hi all,
I have 3 proxmox servers, each with a 4*10G ethernet adapter. I also have two 10G switches to connect them to. The nodes have disks on them to implement a ceph storage and they will also host VMs. I would like to have redundancy in case of a switch failure/reboot, for storage, cluster and VMs.
So:
- For ceph storage the manual says "Bonding can be used to increase availability". Is bonding the only option for redundancy?
- For the cluster I guess I should use ring0 and ring1 for redundancy. Right?
- Should the VM traffic be on the cluster interfaces and let the storage traffic be alone on dedicated interfaces?
- Is there another recommended setup/idea for this config?
Thanx,
Spiros
I have 3 proxmox servers, each with a 4*10G ethernet adapter. I also have two 10G switches to connect them to. The nodes have disks on them to implement a ceph storage and they will also host VMs. I would like to have redundancy in case of a switch failure/reboot, for storage, cluster and VMs.
So:
- For ceph storage the manual says "Bonding can be used to increase availability". Is bonding the only option for redundancy?
- For the cluster I guess I should use ring0 and ring1 for redundancy. Right?
- Should the VM traffic be on the cluster interfaces and let the storage traffic be alone on dedicated interfaces?
- Is there another recommended setup/idea for this config?
Thanx,
Spiros