Skip to content
xaweho

Knowledge base · basic

Configuring a backup job in Proxmox VE — schedule, retention, encryption

How to set up an automatic backup job that regularly sends your VMs and containers to the PBS.

basic ·

A backup job in Proxmox VE is a recurring task that backs up selected VMs and containers to the PBS at the time you set. You create it once, then it runs on its own.

Creating the backup job

  1. In PVE: Datacenter (not an individual VM) → Backup.
  2. Click Add.

In the dialog, you fill in:

“General” tab

FieldRecommendation
StorageYour PBS storage (e.g. xaweho-pbs)
Schedule02:00 for a nightly backup at 2 a.m.
Selection modeInclude selected VMs (or All for everything)
Send emailAlways or On failure only
EmailYour email address
Compressionzstd
ModeSnapshot (for running VMs)
Enableenabled
Job-IDleave empty (generated automatically)

“Retention” tab

This is where you define how many backups are kept. More details in Configuring retention policies.

Recommendation for most setups:

FieldValue
Keep last3 (the last three backups, regardless of when)
Keep daily7
Keep weekly4
Keep monthly6
Keep yearly2

That gives you roughly 22 snapshots per VM, covers daily/weekly/monthly/yearly and uses PBS deduplication efficiently.

”Note Template” tab

Optional. Here you can attach a note to each backup — something like {{guestname}} {{node}} {{job-id}}. Handy when browsing in the PBS GUI.

”Advanced” tab

In most cases, keep the defaults. Only interesting for special requirements:

  • Bandwidth limit (KB/s): if you don’t want to max out your internet connection during the day (default: unlimited at night).
  • PBS Fingerprint: to override the storage’s fingerprint.

Click Add. The job is created and runs automatically at the next schedule trigger.

Schedule syntax

The schedule field uses a syntax similar to systemd.timer. Some examples:

  • 02:00 — daily at 2 a.m. (a time without a weekday means: every day)
  • mon 02:00 — every Monday at 2 a.m.
  • */6:00 — every 6 hours on the hour (so 0:00, 6:00, 12:00, 18:00)
  • mon..fri 02:00 — Monday through Friday at 2 a.m.
  • 02:00,14:00 — daily at 2 a.m. and 2 p.m.

Typical pitfall: keywords like daily or weekly only work on their own and cannot be combined with a time — the parser rejects daily 02:00.

Recommendation for private customers: 02:00 — fits most workloads, runs at night without any performance impact.

Multiple backup jobs for different VMs

You can create as many jobs as you like. Useful, for example:

  • Important VMs: daily, long retention (databases, mail).
  • Standard VMs: daily, standard retention.
  • Test VMs: weekly, short retention.

Per job, you pick the specific VMs in the “General” tab under Selection mode → Include selected VMs.

Testing the backup job

You don’t have to wait until 2 a.m.:

  1. Datacenter → Backup → select the new job.
  2. Click Run now.

The job starts immediately, and you can watch the logs live.

What you should see in the logs

After a successful job:

INFO: Starting Backup of VM 100 (...)
INFO: Backup of VM 100 finished (...)
INFO: Starting Backup of VM 101 (...)
INFO: Backup of VM 101 finished (...)
INFO: Backup job finished successfully
TASK OK

If a single VM backup fails, the job still continues for the other VMs — the job status then becomes “TASK ERROR”.

Frequently asked questions

What happens if a backup doesn’t fit into the daily time window? The backup runs to completion anyway — even if that takes several hours. If the next backup job starts while the current one is still running, the new one waits (PVE serializes backups by default).

How do I see whether all backups ran? In PVE: Datacenter → Backup → select job → Job Detail. There you’ll find the latest task list with success/failure. Alternatively: mail notifications, then you get an email on failure.

Can I run backups with bandwidth limits? Yes, in the “Advanced” tab → Bandwidth limit. Value in KB/s, so 10000 = 10 MB/s = 80 Mbit/s. Handy if the family is still watching Netflix at night.

Are containers (LXC) backed up too? Yes. Backup jobs back up VMs and LXC containers. In the PBS datastore, each backup lands under vm/<id>/... (for VMs) or ct/<id>/... (for containers).

What do Mode: Snapshot vs Suspend vs Stop mean?

  • Snapshot: the VM keeps running, the backup uses a disk snapshot. Default, works for most setups.
  • Suspend: the VM is briefly paused. Only needed on storage without snapshot support (such as local LVM without thin provisioning).
  • Stop: the VM is shut down, backed up, started again. Longest downtime, but the most consistent backups.

In most cases, Snapshot is enough. For critical databases, additionally run pg_dump/mysqldump into a safe folder — then the DB state is guaranteed consistent too.

What’s next

Related products
Tags
pbs proxmox-ve backup-job schedule

Did this article help?

If not, open a ticket. If it did, we're happy about a referral — both sides get €25 credit on their customer account.