Depending on the size of your repository and shutdown window available, you need to plan your go -live and cutover plan very carefully.
Go-live plan - Go-live plan should capture each and every activity in hours and days. The high-level categories that are recommended are
Cutover should be planned for each minute, since shutdown times are always limited. Owners should be clearly mentioned against each activity.
It should typically comprise of -
Go-live plan - Go-live plan should capture each and every activity in hours and days. The high-level categories that are recommended are
- Pre-requisite phase
- Notifying editors
- Cloning of unique instances
- Run maintenance activities etc
- Upgrade phase
- Upgrade author
- Upgrade publish
- Upgrade dispatcher
- Post upgrade phase
- Undertake post-upgrade activities
- Undertake post upgrade performance tunings
- Undertake post upgrade validations
- Build code
- Configure dispatcher with DNS and sling mappings
- Configure ELB
- Setup Akamai CDN
- Configure cold standby
- Backup instances
- Run the sanity test
- Delta content migration
- Get a production clone
- Run the delta content migration using it
- verify delta content
- Verification w/o ELB
- Run sanity validation on each site
- Cutover
- Run delta content migration again (if system was not shut down for above steps)
- Copy it to all clones
- Validate delta content
- Enable ELBs with multiple nodes
- Switch DNS to production
- Go Live
Cutover should be planned for each minute, since shutdown times are always limited. Owners should be clearly mentioned against each activity.
It should typically comprise of -
- Notifying production downtime
- Configure ELB to have one publish
- Shutdown cold standby
- Shutdown all production systems
- Run delta content migration
- Configure runmodes and other configuration needed
- Switch production sling mappings
- Switch dispatcher to production DNS
- Deploy Replication agents
- Configure to have all publish in ELB
- Configure CDN
- Final validation
Comments
Post a Comment