Once the Assessment /Discovery phase is complete and you have the understanding of the current system, its time to do the upgrade.
During the dry run upgrade in the assessment phase, it's important that dry run should be done on production replica, to estimate and plan things during upgrade execution phase.
In our case, there were multiple authors and multiple unique publish instances that we had to upgrade (discovered during assessment phase). And upgrade should be done on high-end servers (we leveraged server with16 core CPU, 70 GB RAM, and high IOPS disk mount), to expedite the upgrade duration.
Once the upgrade is done, its should be deployed on Dev servers and required code base should be built over it.
We used dev server for an upgrade and unit testing. Then test machine was used for integration and QA, UAT (business testing) and LT was used for load testing (dev copy can be leveraged here).
Finally, in the production upgrade, we upgraded again for pre-prod and did a mock cutover on it. Once the sanity was done, same was moved to Production servers.
Before upgrade execution, you should have custom upgrade steps for your project both for authors and publish in somewhat following manner
- Upgrade steps for authors
- Upgrade steps for publish
- Upgrade steps for Dispatcher
- Upgrade troubleshooting
Comments
Post a Comment