Hi everyone. I'm looking for some advice on how I should set up my storage and backups. I'll first share what hardware I currently have and what I need to be able to do with it.
This is what I have as for storage. I've got 4 disks (+ 1 small SSD for proxmox) per node, each 2 matching disks configured as a ZFS mirror for redundancy.
Physically the blades look like this:
Node1: 500GB SSD 500GB SSD 1TB SSD 1TB SSD
Node2: 500GB SSD 500GB SSD 2TB HDD 2TB HDD
Node3: 500GB SSD 500GB SSD 2TB HDD 2TB HDD
Each node has 128GB of RAM, while I'll need about 80GB for VMs per node (probably less). There should be more than enough left for ZFS.
What I want to be able to do is have some containers running on one node, and use 2 nodes to each host a big KVM VM. If something goes down that's not a big issue, but I should be able to recover it on another node in less than an hour. My ZFS mirrors should already protect against disk failure.
It seems I need to be able to use replication between nodes, as I don't have actual shared storage available. Because of the same reason HA is probably not a good idea, but it's also not strictly required.
I've currently configured part of this setup (2 nodes) but already I can't migrate a container between the two nodes because my storage names don't match? How should I set this up, do you have any advice/suggestions?
I have 2 10Gb connections per node, one for cluster communication and one for the rest.
This is what I have as for storage. I've got 4 disks (+ 1 small SSD for proxmox) per node, each 2 matching disks configured as a ZFS mirror for redundancy.
Physically the blades look like this:
Node1: 500GB SSD 500GB SSD 1TB SSD 1TB SSD
Node2: 500GB SSD 500GB SSD 2TB HDD 2TB HDD
Node3: 500GB SSD 500GB SSD 2TB HDD 2TB HDD
Each node has 128GB of RAM, while I'll need about 80GB for VMs per node (probably less). There should be more than enough left for ZFS.
What I want to be able to do is have some containers running on one node, and use 2 nodes to each host a big KVM VM. If something goes down that's not a big issue, but I should be able to recover it on another node in less than an hour. My ZFS mirrors should already protect against disk failure.
It seems I need to be able to use replication between nodes, as I don't have actual shared storage available. Because of the same reason HA is probably not a good idea, but it's also not strictly required.
I've currently configured part of this setup (2 nodes) but already I can't migrate a container between the two nodes because my storage names don't match? How should I set this up, do you have any advice/suggestions?
I have 2 10Gb connections per node, one for cluster communication and one for the rest.