OpenVTL™ help — first-time operator guide

Applies to v1.0.0. A guided path from a blank Debian box to an IBM i saving to virtual tape with offsite copies — we ran every step below on a fresh appliance and a real IBM i host over 8 Gb FC during v1.0.0 release validation. Screenshots show a demo appliance (demo01, libraries VTL01/VTL02); identifying values (WWPNs, keys, addresses) are illustrative.

Read in order the first time:

  1. Install the appliance — bundle → running web UI, one reboot.
  2. First storage and library — disks → pool → library → cartridges.
  3. Connect the IBM i host — ACLs on the appliance, DEVD/WRKMLBSTS on the host.
  4. S3 offsite — add a remote, export, evict, re-import.
  5. Recovery — single-cart restores and one-click whole-library DR.

Day-2 operations (boot recovery, full pools, updates, incident index) live in the operator runbook (docs/operator-runbook.md, shipped inside every release bundle). This guide is the happy path; the runbook is what you read when something is off.

The five-minute mental model

  • One Go daemon, openvtld, is the control plane: web UI on https://<host>:8443, everything admin-gated and audited. The data path (mhVTL tape emulation + LIO FC target) runs independently — a stopped control plane never interrupts a running backup.
  • Tape data lands on ZFS (zstd + global dedupe, dedupe table on a dedicated SSD). A pool is a dataset; a library pairs with exactly one pool.
  • The host sees a real IBM library — TS3100/TS3200 (3573) or TS3500 (3584) with IBM LTO drives. Identity (serials, NAA) survives reboots and rebuilds, so BRMS never re-enrolls anything.
  • Vaulted cartridges tier to any S3-compatible bucket; the bucket alone is enough to bring a dead library back (labels preserved).
Start1 · Install the appliance →