Proxmox Backup Server (PBS)¶

Proxmox Backup Server running on px5-lemans (France) for efficient incremental backups over WAN.

Container: CT 5101 (pbs-fr)
IP: 10.35.1.101
Web UI: https://10.35.1.101:8007
Last Updated: 2026-02-04

Overview¶

PBS replaces traditional vzdump-to-NFS backups with:

Incremental backups - Only changed data transferred (typically 1-5GB vs 40-100GB)
Deduplication - ~50-70% storage reduction
Resume on failure - Interrupted backups continue where they left off
Faster restores - 10-20 minutes vs 30-60 minutes

Architecture¶

UK (px1/px2/px3)                              France (px5)
      │                                            │
      ├── PBS ────── WireGuard (~18ms) ──────→ CT 5101 (pbs-fr)
      │              (incremental, ~1-5GB/night)   │
      │                                            ↓
      │                                       PBS Datastore
      │                                       (pikvm-backup NFS)
      │
      └── vzdump → px3-nas (UK local, fast restore)

Container Specification¶

Parameter	Value
VMID	5101
Hostname	pbs-fr
IP	10.35.1.101/24
Gateway	10.35.1.1
Cores	4
RAM	4096 MB
Root disk	16 GB (local)
Datastore	Bind mount from pikvm-backup NFS

Datastore¶

Name	Path	Storage	Free Space
pbs-main	/mnt/datastore/pbs-main	pikvm-backup NFS	~1.1 TB

Maintenance Jobs¶

Job	Schedule	Purpose
Garbage Collection	Saturday 02:00	Remove orphaned chunks
Verification	Sunday 04:00	Verify backup integrity

PVE Storage Integration¶

PBS is registered as storage pbs-fr in the Proxmox cluster.

# Check storage status
pvesm status | grep pbs-fr

# List backups
pvesm list pbs-fr --content backup

Backup Jobs¶

All jobs configured in /etc/pve/jobs.cfg:

Job	Node	Schedule	Retention
pbs-px3-daily	px3-suzuka	22:00	7 daily, 4 weekly, 2 monthly
pbs-px2-daily	px2-monza	00:30	7 daily, 4 weekly, 2 monthly
pbs-px1-daily	px1-silverstone	03:00	7 daily, 4 weekly, 2 monthly
pbs-weekly	px1-silverstone	Sun 07:00	8 weekly, 3 monthly

Job Details¶

vzdump: pbs-px1-daily
    all 1
    compress zstd
    enabled 1
    mailnotification failure
    mailto chpa35@gmail.com
    mode snapshot
    node px1-silverstone
    notes-template {{guestname}}-pbs
    prune-backups keep-daily=7,keep-weekly=4,keep-monthly=2
    schedule 03:00
    storage pbs-fr

Web UI Access¶

URL: https://10.35.1.101:8007

Warning

Accept the browser security warning for the self-signed certificate.

Realm: Linux PAM
Username: root
Password: Stored in password manager

Key Dashboards¶

Dashboard - Overview of datastore usage and recent tasks
Datastore > pbs-main - Browse backups, verify, prune
Administration > Tasks - View running and completed jobs

CLI Commands¶

On PBS Container (CT 5101)¶

# Enter container
ssh px5 "pct exec 5101 -- bash"

# List datastores
proxmox-backup-manager datastore list

# Show datastore details
proxmox-backup-manager datastore show pbs-main

# List users
proxmox-backup-manager user list

# Check GC status
proxmox-backup-manager task list --type gc

# Manual GC
proxmox-backup-manager garbage-collection start pbs-main

From PVE Nodes¶

# List PBS backups
pvesm list pbs-fr --content backup

# Manual backup to PBS
vzdump 1112 --storage pbs-fr --compress zstd --mode snapshot

# Restore from PBS
pct restore 1112 pbs-fr:backup/ct/1112/2026-02-04T03:00:00Z --storage ceph-pool

# Restore to different VMID (test)
pct restore 9999 pbs-fr:backup/ct/1112/2026-02-04T03:00:00Z --storage ceph-pool --unique

Recovery Procedures¶

Restore CT from PBS¶

# List available backups
pvesm list pbs-fr --content backup | grep "ct/1112"

# Restore (overwrites existing)
pct restore 1112 pbs-fr:backup/ct/1112/2026-02-04T03:00:00Z --storage ceph-pool

# Restore to new VMID
pct restore 9999 pbs-fr:backup/ct/1112/2026-02-04T03:00:00Z --storage ceph-pool --unique

Restore VM from PBS¶

# List available backups
pvesm list pbs-fr --content backup | grep "vm/1111"

# Restore
qmrestore pbs-fr:backup/vm/1111/2026-02-04T03:00:00Z 1111 --storage ceph-pool

File-Level Restore¶

PBS supports restoring individual files without full VM restore:

Open PBS Web UI
Navigate to Datastore > pbs-main > Content
Select backup snapshot
Click "Browse Files"
Download individual files

Monitoring¶

Check Backup Status¶

# Recent PBS tasks
ssh px5 "pct exec 5101 -- proxmox-backup-manager task list" | head -20

# Check datastore usage
pvesm status | grep pbs-fr

# Verify backups exist
pvesm list pbs-fr --content backup | wc -l

Alerts¶

PBS sends email notifications on: - Backup failures - Verification failures
- Low storage warnings

Troubleshooting¶

PBS Unreachable¶

# Check container status
ssh px5 "pct status 5101"

# Check PBS services
ssh px5 "pct exec 5101 -- systemctl status proxmox-backup-proxy"

# Check network
ping 10.35.1.101

Storage Not Visible¶

# Check PVE storage config
grep -A10 "pbs: pbs-fr" /etc/pve/storage.cfg

# Test connectivity
pvesm list pbs-fr --content backup 2>&1 | head -5

Backup Failures¶

# Check vzdump logs
tail -100 /var/log/vzdump/vzdump-*.log

# Check PBS task log (in web UI or CLI)
ssh px5 "pct exec 5101 -- proxmox-backup-manager task list"

Parallel Run Period¶

During the 2-week validation period (Feb 4-18, 2026):

Both vzdump and PBS jobs run simultaneously
Monitor PBS backup sizes (should decrease after day 2)
Test restores from PBS
Compare with vzdump restore times

After validation, disable old vzdump jobs: - pikvm-px1, pikvm-px2, pikvm-px3 - backup-55292acc - weekly-archive

Keep active: uk-secondary (vzdump to px3-nas for fast UK restores)

Expected Benefits¶

Metric	Before (vzdump NFS)	After (PBS)
Nightly WAN transfer	40-100 GB	1-5 GB
Storage used	~1.5 TB	~500 GB
Restore time (France)	30-60 min	10-20 min
Resume on failure	No	Yes
Deduplication	No	Yes

Last updated: 2026-02-04