Avoiding Sync Job Errors with Rotating Offsite Datastores in Proxmox Backup Server

[CS_Luca]

I'm running into an issue with my Proxmox Backup Server setup and now i need some help.

I have two external offsite drives configured as separate datastores. I rotate them manually, meaning only one of the two disks is connected to the server at any given time. Each datastore has its own sync job configured.

The problem: whenever one disk is disconnected (which is intentional), the corresponding sync job fails and triggers an email notification ("Disk is not mounted"). This results in constant "false positive" error alerts, even though everything is working as intended with the currently connected disk.

What I’m looking for is a clean solution where:
1. The sync job for the disconnected datastore does NOT throw an error
2. Only the active/connected datastore runs its sync job normally
3. Ideally, no unnecessary failure emails are sent when a disk is simply not present

Has anyone solved this in a smart way? Maybe conditional job execution, hooks, or a different datastore strategy?

Would appreciate any ideas or best practices

 

[d.oishi]

You may want to try the sync job options --run-on-mount and --unmount-on-done. Both options are documented in the official CLI documentation.
https://pbs.proxmox.com/docs/command-syntax.html#proxmox-backup-manager

If you enable these options and leave the Sync Schedule unset (remove the schedule), you can use the mount event of the removable datastore as the trigger for the sync job. When run-on-mount is enabled, the sync job is started automatically whenever the relevant removable datastore is mounted. If unmount-on-done is enabled, the datastore can be unmounted automatically after the sync job has finished.

With schedule-based sync jobs, PBS will still start the sync even if the removable datastore is not mounted, and the job will fail. For rotating offsite disks, a run-on-mount based workflow is therefore probably a better fit than a calendar-based schedule.