Backup strategy for the Plesk vServer

intermediate · Last updated 16 July 2026

Backup on a dedicated vServer has more layers than on shared hosting. Plesk backup, VM-level server snapshots, plus optional cloud backups of your own. Here’s what runs automatically on our side and what you can add yourself.

What we do automatically

Plesk daily backup (every 6 hours)

Standard on the vServer: a Plesk backup every 6 hours, kept for 14 days.

Contents:

Website files
Mailboxes
Databases
Plesk config (service plans, end customers, etc.)

Four backup states per day — if something goes wrong in the morning, you can roll back to 6 p.m. yesterday.

VM snapshots at hypervisor level

On top of that, we take server snapshots of the entire VM:

Daily snapshots (7 days back)
Weekly snapshots (4 weeks back)
Monthly snapshots (3 months back)

VM snapshots include:

The complete OS
Plesk + all apps
Custom config (added by us or through tickets)

If the Plesk backup doesn’t reach back far enough or the OS itself is broken: the VM snapshot is the lifeline.

On hardware failure: VM migration

RAID10 protects against individual disk failures. For larger hardware problems:

The VM is migrated to another host (typically live, sometimes with brief downtime)
Your data comes along 1:1
Best case you notice nothing, worst case 5–15 minutes of downtime

Self-service restore from the Plesk backup

In Plesk → Account → Backup & Restore:

List of backups (4 per day, 14 days = ~56 snapshots)
Click a backup → Explore contents → find the individual file → restore
Or a full restore

Works for:

A single file of an end-customer site
An entire end-customer account
Your whole Plesk area (aborts if there are active sessions)

Restore duration: small restores take minutes, a full restore takes hours.

More in Creating backups and restoring (Plesk Web Hosting KB) — the mechanics are identical.

Restore from a VM snapshot

VM snapshots aren’t directly visible in Plesk — we restore from them via ticket.

When VM snapshot instead of Plesk backup?

Plesk backup is broken / doesn’t reach back far enough
OS or Plesk stack is broken (not fixable via Plesk backup)
Custom config lost — the Plesk backup doesn’t include it
A major update (Plesk or OS) wrecked the stack

How to request a restore

Ticket to us:

Please restore the server snapshot from . Reason: the Plesk update on broke php-fpm, sites are returning 502.

We:

Verify snapshot availability
Agree a time slot with you (brief downtime required)
Perform the restore
Test, then confirm back to you

Duration: restore time 30-60 minutes, plus coordination.

Your own backup strategy on top

If our standard isn’t enough for you:

Plesk backup to an external target

In Plesk → Account → Backup & Restore → Remote Storage:

FTP/SFTP to your own server
Amazon S3 / Backblaze B2
Google Cloud Storage
WebDAV

Plesk pushes an additional backup to that target.

Use case: you want a backup under your own control, not only at xaweho.

Manual DB dumps to your own storage

Via cron job (configurable in Plesk):

mysqldump --all-databases | gzip > /backup/db-$(date +%Y%m%d).sql.gz
rsync -avz /backup/ user@dein-server:/path/

Specifically for DB-heavy setups.

restic / borg for files

Via SSH (chrooted):

restic backup ~/  --repo sftp:user@dein-server:/repos

Open source, encrypted, deduplicated.

Off-site backup to a Synology at a second location

We can provide a backup Synology at one of our other locations (Nuremberg, Falkenstein) for Plesk to push to.

Advantage: off-VM, off-hypervisor, at a second location. Extra charge: similar to the Synology backup add-on (see the storage plans).

Open a ticket if you want this.

Backup tests

Important: a backup without a successful restore test isn’t really a backup.

Quarterly test restore

Recommended every 3 months:

Restore a random file from the Plesk backup
Check that the content is correct
Check that encryption passwords still work

Disaster recovery drill

Once a year (or after major changes):

Play through a hypothetical “server completely gone” scenario
How would you request the restore?
What would the critical time window be?

We help with drill planning — a ticket is all it takes.

What’s not in the backup

Other vServers (separate setups, separate backups)
External data (Cloudflare cache, external APIs)
Live services holding data in that moment (Memcached, Redis)
Mail in the mail clients of your end customers (that lives on their devices)

The role of RAID10 mirroring

RAID10 is NOT a backup. It’s resilience against disk failure:

One disk dies → no downtime, no data loss
Two disks in the same mirror group die → data gone

But: RAID10 does not protect against:

Accidental deletion
A software crash with data corruption
A hack
Ransomware

Hence: RAID10 + Plesk backup + VM snapshot — three layers, each against a different risk.

Recommended strategy for vServer customers

Standard setup (for most people)

Plesk backup (every 6h, 14 days) — standard, we handle it
VM snapshots — standard, we handle it
RAID10 — standard, we handle it

That covers 95% of all cases. Nothing extra needed.

Extended setup (compliance, critical data)

Plus Plesk backup to an external target (your own or our Synology)
Plus regular DB dumps to separate storage
Plus off-site backup at a second location

A bit more effort, higher safety.

Premium setup (business-critical)

Everything from the extended setup
Plus automatic snapshot replication to a second location (on request)
Plus a hot-standby server for immediate failover (custom setup, write to us)

Frequently asked questions

How big is the Plesk backup footprint? Compressed, ~50-70% of the original data. With 100 GB of live data: ~70 GB of additional backup storage, multiplied by 14 days of retention.

For large setups: backup disk costs more than live disk. That can justify a plan upgrade.

Does backup storage count against my plan quota? Yes. So for large sites: an external backup target makes sense — it takes load off the server disk.

Encrypting backups — yes or no? We recommend yes. Server-side in Plesk: optionally on. Keep the backup password somewhere safe.

How long does a restore from a VM snapshot take? 30-90 minutes, depending on server size.

Can end customers do self-service restores? If the service plan allows it — yes, of their own files. Server-wide restores are done only by us.

What if an end customer loses data? The end customer does a self-service restore from their Plesk backup tab. If that’s not available: you as the reseller do the restore. If that’s not enough either: VM snapshot via us.

How old can a backup be that I restore? Plesk backups: up to 14 days back. VM snapshots: up to 3 months back. Older: only if you keep your own external backup.

Restore from another vServer onto this one — possible? Cross-server restore is possible, with the Plesk Migrator or via ticket. Watch out for DNS/SSL config — it may need adjusting.