pbs-ana

2026-10-04: PBS 4.2.7 on Debian 13.7 (kernel 6.12.111), upgraded from 3.4.8 (docs/runbooks/pbs-pbs4-upgrade-plan.md). PBS-on-plain-Debian install (no proxmox-backup meta package, Debian kernel), so pbs3to4 FAILs one check by design. infra-ops has no login here: root ops go through the hypervisor's guest agent (qm guest exec).

Proxmox Backup Server — fleet primary. VM on pfi-pve with an NFS-mounted datastore on the Ana NAS. All 5 hypervisors (pfi-pve, nh3-pve, esh-pve, esh-pve-nas, sfsrv-ana) back up here; one-way sync to pbs-nh3 at 06:00 daily.

Network

LAN IP: 10.250.50.90
FQDN: pbs-ana.phasefinal.com
PBS web UI: https://10.250.50.90:8007/
SSH: ssh pbs-ana (config alias → lkraven@10.250.50.90). Root ops via sudo from lkraven.
Homepage card: PBS-ANA in Infra - ANA group.

Hardware (VM)

Hypervisor: pfi-pve
vCPU: 4 (QEMU)
RAM: 7.8 GB
Root disk: /dev/sda1 31 GB ext4 — OS only (2 GB used)
Datastore: NFS from ana-nas at /mnt/pbs-datastore → 10.250.50.50:/mnt/backup/pbs-ana, 20 TB (currently 336 GB used)

NFS details (the ZFS-case-insensitivity gotcha)

The ZFS dataset backing /mnt/backup on ana-nas is casesensitivity=insensitive, which breaks NFSv4 writes (EACCES). Mount must be NFSv3 — that's what /etc/fstab already uses:

10.250.50.50:/mnt/backup/pbs-ana /mnt/pbs-datastore nfs \
  defaults,_netdev,bg,hard,timeo=600,retrans=2,vers=3 0 0

Don't upgrade this to vers=4 without fixing the dataset. See docs/runbooks/pbs-deployment.md Phase 0.5 for the full history.

What runs here

proxmox-backup-proxy — HTTPS API + web UI (port 8007)
proxmox-backup — backend service
Scheduled verify job: weekly Saturday 23:00 (ignore-verified=true, outdated-after=30 days)
Garbage collection: default weekly

Namespaces

Per-hypervisor namespaces to prevent VMID collisions across PVE hosts (VMIDs are NOT globally unique):

pfi-pve
nh3-pve
esh-pve
esh-pve-nas
sfsrv-ana

API token (for hypervisor clients)

root@pam!fleet-vzdump — stored in password manager; referenced by each hypervisor's /etc/pve/storage.cfg pbs-ana stanza.

Sync to DR

One-way pull: pbs-nh3 pulls from here at 06:00 daily. pbs-ana does not push; all coordination is on the DR side.

Dependencies / blast radius

Hard dependency on ana-nas (CT 109). Datastore NFS goes dead if ana-nas crashes. Fleet-wide vzdump + verify + sync all fail until NFS recovers. See memory/incident_ana_nas_spof.md.
Self-backup excluded from scheduled vzdump — backing up a PBS VM to itself is a circular dependency. Config-only backup could be captured via file-level restic if needed; not currently set up (low priority — the datastore is on NFS, OS is reinstallable in ~30 min).

Refresh

scripts/refresh-server-info.sh pbs-ana

Snapshot at system-details.txt. Full deployment history: docs/runbooks/pbs-deployment.md.