I currently have two Mini PC with Proxmox in a cluster along with a raspberry Pi QDevice.
Each server has a public 2.5gbe and a private 2.5gbe (interconnected between the servers on separate subnet).
I am using ZFS on NVME on both servers.
I am currently using HA w/ Replication.
1) Is replication going to be the best way to prevent data loss and allow for fail over when a node goes down? I know CEPH is out of the question with only two nodes, is GlusterFS a better choice?
2) Will replication honor the cluster NIC priority, and use the interconnected 2nd nic for storage traffic along with cluster traffic when possible? I have the interconnected nic as the higher priority.
3) Is watchdog based fencing my best option here?
Each server has a public 2.5gbe and a private 2.5gbe (interconnected between the servers on separate subnet).
I am using ZFS on NVME on both servers.
I am currently using HA w/ Replication.
1) Is replication going to be the best way to prevent data loss and allow for fail over when a node goes down? I know CEPH is out of the question with only two nodes, is GlusterFS a better choice?
2) Will replication honor the cluster NIC priority, and use the interconnected 2nd nic for storage traffic along with cluster traffic when possible? I have the interconnected nic as the higher priority.
3) Is watchdog based fencing my best option here?