Backup to 2nd (cluster) PVE box

[Thomas Stone]

Hello:

I want to backup VMs from machine 1 to machine 2. If I create a storage location (type, folder) on machine 2 and add it as storage for VZDump Backup Files that are accessible by both machines, when the backup runs, it does not go to the actual disk on the machine 2.

On machine 1 the exact same backup path is created above /, so like /diskname/path. But what I wanted was to reference the zfs disk (filesystem) referenced on machine 2.

I'm a noob in the cluster department, so probably something obvious. I think I need to know how to set the path/target so it references machine 2.

So, is there any way to add storage to a cluster that is on machine 2 and run a backup of a VM on machine 1 so the data actually goes to the disk on machine 2?

I do not have shared storage separate from either of the PVE boxes and I'm not trying to do HA. I use ZFS on all disks.

Thank you!

 

[proxtest]

If u want distribution und need shared storage! Proxmox dont create this for u.
Use drbd or iscsi or nfs to share a storage.

 

[LnxBil]

Easiest thing would be to create a NFS-Export on your machine2 and mount it in machine1 as backup storage. Then the backup will automatically go there.

 

[Thomas Stone]

Thanks for the replies. Actually, pve-zsync is going to be useful for me. Machine 2 will only boot and run for a few hours daily to do the updates. So, is NFS going to be tolerant of the device going on/offline regularly (and auto mounting on Machine 1 every time it comes up) or do you recommend a different protocol to share the drives in Machine 2?

Thanks!

 

[LnxBil]

Thanks for the replies. Actually, pve-zsync is going to be useful for me. Machine 2 will only boot and run for a few hours daily to do the updates.

That is also a good solution ... or even a better one, yet snapshots are no backups.

So, is NFS going to be tolerant of the device going on/offline regularly (and auto mounting on Machine 1 every time it comes up) or do you recommend a different protocol to share the drives in Machine 2?

No, NFS is very intolerant to network failures and should not be used in your scenario. It has to be there all the time. pve-zsync is your way to go on ZFS. I missed that part that you're running ZFS on both nodes.