Install the new bare metal server for main postgresql replacement

Orders: https://mybox.inria.fr/smart-link/0c2a300a-559a-40b5-9978-c6e086935872/

Inventory: https://inventory.internal.admin.swh.network/dcim/devices/264/

Summary:

• Management address (DNS): 128.93.134.49 (swh16-adm.inria.fr)
• VLAN configuration: VLAN440
• Internal IP: 192.168.100.212
• Internal DNS name(s)[1]: albertina.internal.softwareheritage.org

Tasks:

• Declare the servers in the inventory
• Add the management info in the credential store
• Upgrade firmwares
• Install the OS
• Add puppet configuration (if needed)
• Configure firewall rules (if needed)

[1] https://intranet.softwareheritage.org/wiki/Hostname_naming_scheme

Database migrations:

• indexer
  • stop indexer journal clients
  • finalize database transfer
  • pg_upgrade indexer database to 16.x on albertina
  • start database on albertina
  • migrate indexer storages to new database
• main storage database
  • stop background loaders
  • stop add forge now loaders
  • move swh-web to secondary database
  • stop deposit loaders
  • notify user-visible save code now suspension
  • stop save code now loaders
  • finalize database transfer
  • pg_upgrade storage database to 16.x on albertina
  • start database on albertina
  • migrate postgresql storages to new database
  • restart save code now loaders
• secondary database
  • stop listers
  • stop scheduler runners
  • stop scheduler journal client
  • stop scrubbers
  • stop vault workers
  • stop vault backend
  • notify of user-visible swh-web (authenticated) / swh-vault / swh-deposit outage
  • stop deposit backend
  • stop swh-web
  • finalize database transfer
  • pg_upgrade secondary database to 16.x on albertina
  • start database on albertina
  • migrate postgresql secondary references to new database
• finish user-visible maintenance
  • restart swh-web
  • restart deposit backend
  • restart deposit loaders
  • restart vault backend
  • restart vault workers
  • restart scheduler runners
  • notify end of user-visible maintenance
• finish maintenance window
  • restart scheduler journal client
  • restart add forge now loaders
  • restart listers
  • restart background loaders
  • restart indexer journal clients
  • restart scrubbers
  • close maintenance window