Docs ·

Supfile operations runbook

Control sequence for 17 VMs across three sites.

Initial checks

  1. Verify Tailscale connectivity for every VM.
  2. Check HTTP health in Paris and New York.
  3. Inspect wsrep state on all three Galera nodes.
  4. Confirm that all six GlusterFS bricks are online.

Failover test

Record Prometheus state, stop one web node, observe local recovery, then test site loss. Preserve every result as evidence.