Restore Playbooks

This page collects “when to use what” restore examples. For the full explanation of modes and safety rules, see docs/docs_file/restore_modes.md.

Before you restore

• Restores require root and write to /.
• ProxSave never overwrites /etc/pve; those files are exported instead.
• Prefer –cli –log-level debug if you are troubleshooting.

Playbook 1 — Fix a broken network/SSH/service layer (SYSTEM BASE)

Use when the node boots and disks are fine, but access/services are misconfigured.

proxsave --restore --cli

Choose SYSTEM BASE (base) and restore categories such as network, ssl, ssh, services.

Playbook 2 — Restore storage/datastores and jobs (STORAGE/DATASTORE)

Use when PVE storage.cfg / PBS datastore configuration and job config are broken, but you don’t want a full node restore.

proxsave --restore --cli

Choose STORAGE/DATASTORE (storage).

Playbook 3 — Selective “surgical” restore (CUSTOM)

Use when you want to restore only a couple of categories (e.g., only ssl + services).

proxsave --restore --cli

Choose CUSTOM and toggle only the categories you want.

Playbook 4 — Full rebuild after OS reinstall (FULL)

Use when you rebuilt the OS and want to restore most configuration from a backup bundle.

proxsave --restore --cli --log-level debug

Choose FULL (full) and carefully review the restore plan before confirming.

Playbook 5 — PVE cluster disaster recovery (SAFE vs RECOVERY)

If ProxSave detects a cluster bundle and your restore plan includes pve_cluster, it prompts:

• SAFE: export cluster files, do not write /var/lib/pve-cluster/config.db
• RECOVERY: restore the cluster database (use only when the cluster is offline/isolated)

Start restore:

proxsave --restore --cli --log-level debug

Related docs:

• docs/docs_file/cluster_restore_for_disaster.md
• docs/docs_file/restore_capabilities_and_classes.md

Where restore logs go

• Detailed restore log: /tmp/proxsave/restore_YYYYMMDD_HHMMSS.log
• Safety backup tarball: /tmp/proxsave/restore_backup_YYYYMMDD_HHMMSS.tar.gz