A reason to build.
Services and storage had outgrown their original host. Consolidating them meant moving real workloads while preserving access, data, and a recovery path.
How the pieces fit.
- 01
Staged the storage and service migration before the network cutover, using a production VM and dedicated containers for different workloads.
- 02
Moved the media-serving workload onto a container with local storage bindings, removing the previous NFS dependency for that path.
- 03
Migrated smart-home services separately and repurposed the former server as an administration workstation.
- 04
Used backups, retained rollback resources, and a verification soak to check the migrated core rather than treating a successful copy as completion.
What came out of it.
Migrated the core storage, production services, media serving, and smart-home workloads. The documented core passed its verification soak; deferred migration items were tracked separately.
Documented staged cutover, backup configuration, workload checks, incident diagnosis, and verification soak. Private infrastructure; the topology here is intentionally abstract.
What I’d carry forward.
A hardware NIC hang looked like a migration failure. Checking the hypervisor and guest separately prevented a bad diagnosis. Hypervisor renames also need their supported migration procedure, not a normal hostname change.
Working on something with similar edges?
Let’s compare notes ↗