I would like to setup two separate Proxmox HA Clusters at two different sites. In case an entire site fails, I need the possibility to restart its VMs at the other site. I guess it should be possible to setup a Proxmox Backup Server at each site, regularly backup the other site's VMs to it and in case one site fails, restore its VMs from the second site's backup server to the second site's HA cluster to start them there. However, this would probably take a quite some time. It would be much nicer if it was possible to regularly sync the backups to the "foreign" cluster so that in case it needs to take over, they can be immediately started. Is something like that possible?
An alternative would be to combine the two clusters into a single one and use replication, but I'm not sure if that would work. First, in each HA cluster, the nodes are connected via multiple separate 25Gbit/s networks (corosync, Ceph), while the two sites are only connected via VPN over the internet (max. 1Gbit/s). I also don't know if replication would even be possible from one Ceph storage to another one (and having only a single Ceph cluster shared between the different sites is probably horribly slow). In addition, I would need to setup a third site for a tie-breaker, and I really don't need all of that. I just want to be able to manually startup a (fairly) recent version of the VMs of the failed site at the other site in a very short time.
An alternative would be to combine the two clusters into a single one and use replication, but I'm not sure if that would work. First, in each HA cluster, the nodes are connected via multiple separate 25Gbit/s networks (corosync, Ceph), while the two sites are only connected via VPN over the internet (max. 1Gbit/s). I also don't know if replication would even be possible from one Ceph storage to another one (and having only a single Ceph cluster shared between the different sites is probably horribly slow). In addition, I would need to setup a third site for a tie-breaker, and I really don't need all of that. I just want to be able to manually startup a (fairly) recent version of the VMs of the failed site at the other site in a very short time.