Migration design
Target architecture, dependency mapping, storage and network design, backup and recovery approach, and the rollback position for every stage.
A VMware exit is a sequencing problem before it is a software problem. The order things move in, what gets tested before anything is committed, and where you can stop and go back — that is what decides whether it goes well.
Target architecture, dependency mapping, storage and network design, backup and recovery approach, and the rollback position for every stage.
A representative set of workloads moves first. Performance, backup and restore are tested on the new platform before anything else is committed.
Workloads move in controlled groups, each with an agreed downtime window and a rollback point that has been checked rather than assumed.
Where it is needed: storage, monitoring, logging, identity, backup and network services. Only the layers you actually run.
Documentation, runbooks and knowledge transfer, so the platform is yours to operate rather than ours to keep.
Virtualisation
Storage
Backup
Monitoring
Logging
Identity
Networking
Object storage
The sequencing, the storage layer, and anything that can take production down.
So they know the system afterwards, not just the documentation.
Not written up at the end from memory.
Each stage leaves you in a working state.