Setup HA collectors
Phase 3: Set up HA collectors
- Verify and debug any puppet issues
- Reinstall the machine to verify final puppet configs
- Install the new service
- Connect to the existing pool, update frontend configs for new backup
- Verify no issues
- Shoot primary HA collector
- Update FE configs to talk to new collector
- Update the FE to the new collector. Turn off negotiator for old pool. Let new pool negotiator take over.
- Once old pool is drained turn off collector.
- Wait some amount of time to verify no issues, can retire the old collector nodes
- Do we also want to do this staged/tested with ITB or can we go directly to production (which already has a backup)? Factory/schedd don’t have this so need to test with ITB first