It's probably complex enough that you'll want to script it.
It's how I set up a highly available, replicated SAN system in two geographic locations a long time ago (half decade ago, so I'm working off memory).įreenas could be set up to automatically squash snapshots on the host system as well as the target system, making disaster recovery faster since you'll have max X snapshots to layer during restoration.Ĭheck the freenas forums and docs to get a high level overview of the approach then check the zfs CLI tool docs to figure out how to do it yourself in the proxmox environment. ago It is not system drive failure I am talking about. Subsequent snapshots are smaller in size and quicker to transfer, and if you have to recover you import each layered snapshot to the new pool one by one. You can use the Proxmox Backup Server client to backup the system drive: In this case, you should definitely be running PBS on a separate machine so you can still access the backups when the system fails. I think there's a certain file format you can save the snapshot as which can be used to build the data of a zpool later during recovery.
The approach freenas takes is make a snapshot then replicate the snapshot to a remote machine. This weekend I'm potentially digging into how freenas does zfs replication to see if the same can be done on proxmox zpools.
0 kommentar(er)
