I recently migrated some of my infrastructure to Proxmox and now run my first 2-node cluster using a qdevice for quorum. Each node is connected to a Synology NAS over NFS. The setup works smoothly overall, except for backups.
With qcow2 disks on NFS, the CBT (changed block tracking/disk bitmap) data is discarded after a VM reboot. As a result, if a VM or an entire node restarts (say due to maintenance or similar) the next backup run can a very long time. I've got over 20 TB of data to backup on this cluster, so this is going to take forever.
Historically, I have used Veeam, but I want to get off that in favor or Proxmox Backup Server. However, since this is a hypervisor/qemu level issue, I don't think whatever software solution I use matters? They will all run into the same issue as the bitmap is erased on VM reboot.
Looking for effective strategies to make backups more sustainable in this setup - has anyone got a practical workaround for this? Proper backups are essential. I'll even settle for a way to speed up the backup.
With qcow2 disks on NFS, the CBT (changed block tracking/disk bitmap) data is discarded after a VM reboot. As a result, if a VM or an entire node restarts (say due to maintenance or similar) the next backup run can a very long time. I've got over 20 TB of data to backup on this cluster, so this is going to take forever.
Historically, I have used Veeam, but I want to get off that in favor or Proxmox Backup Server. However, since this is a hypervisor/qemu level issue, I don't think whatever software solution I use matters? They will all run into the same issue as the bitmap is erased on VM reboot.
Looking for effective strategies to make backups more sustainable in this setup - has anyone got a practical workaround for this? Proper backups are essential. I'll even settle for a way to speed up the backup.