GCVE exit planning

[steveoh]

Benefit

The license change for VMWare has DTS in a pickle with low adoption rates and large license costs. Planning has started with a spreadsheet about what the future of the GCVE machines will be. It's on a 1-5 year schedule. It would be beneficial to start chipping away at our GCVE footprint to make the transition easier.

Acceptance Criteria

• Mapserv is migrated to portal Migrate mapserv services to portal #455 and retired. Requires Portal Collaboration for Address Points #409
• The GCVE NSX-T load balancers are shut down
• The mapserv test server is retired
• All databases on itdb002gp are migrated itdb002sp database dependencies #506
• The TURN GPS network is migrated to the cloud*
• The Nevada GPS network is migrated to the cloud*

If the SGID goes to the cloud

• The forklift machine is shut down
• Forklift runs cloud native
• Auditor runs cloud native
• Conductor runs cloud native
• Sweeper runs cloud native
• Cambiador pointed at new cloud database
• CloudDB pointed at new cloud database
• Red Rocket Roads is migrated to a VM next to the SGID/UTRANS or consider using the CloudSQL Proxy and retire the vm

If the SGID goes on prem

• The forklift machine follows the SGID
• Red rocket moves on prem

If we no longer have on prem servers

• Citrix is shut down

*Requires new IP address for all receivers

Notes

• Consider Microsoft licensing costs when migrating to CloudSQL. PostgreSQL would save a lot of money
• Consider if Enterprise Geodatabases are required or if PostGIS covers the needs
• PRJTASK0637484 can be used for any work related to the GCVE exit project
• Lucid drawing

Risks

• Moving databases will disrupt a lot of users. We may need to mirror our environment for a while to operate on bite size chunks
• Assigning new IP's for the GPS network is a huge pain and will disrupt the business possibly losing customers
• Paying work will not be possible as this is a huge undertaking

Issue Reference

No response

[stdavis]

Conductor and cloudb already runs in the cloud, no? Or perhaps you are just putting them there because they will need to be repointed at cloud databases?

[steveoh]

Good catch. Yes, we will need to rethink the tools in our lifecycle.