We currently don't have automation to cover that particular area.
This, on paper, would mostly be a variation of the clone/copy cookbook, handling a pre-replication set of tasks.
Description
Description
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | Volans | T362893 Spicerack support for dbctl | |||
| Open | None | T369295 Cookbook to create a sanitarium host from scratch |
Event Timeline
Comment Actions
A sanitarium master doesn't require anything special, it is just a normal production slave
Comment Actions
So a quick high level of how this would be:
- Select a host and assign all the needed puppet roles
- Clone it from production (recommended: from its future sanitarium master)
- Including setting replication up
- Move it to replicate under its desired sanitarium master (ideally the host it was recloned from)
- Let it catch up with its master
- Run redact_sanitarium script across all the instances
- Run check_private_data.py to make sure all the PII is filtered (I usually run it once, let a few hours past, and then run it again to make sure triggers, and replication filters are working as expected)