Has anyone been involved or had the experience of doing rolling releases? Essentially, our development team will be working on multiple devlierables and as they finish development, the project will be moved in our QA environment. QA will test the new deliverable and then hold it in the environment as the next deliverables are moved to QA. Once QA has approved all deliverables everything will be moved to an integration environment and then to production.
What are some things I should look out for? Are there extra precautions that should be put in place from the QA team? Any advice (good and bad) you have is greatly appreciated.
Are fixes and enhancements that the development team is working on structured into components? If they can be thus structured, then with each release / movement from the dev, you could do a Component level testing and a Component Assembly testing, instead of testing the system each time.
If you donít have release structured this way, then you are in a bind.
One good thing to do would be for you to establish a functional tree and hierarchy, prioritize functions and establish function level test plans that can be executed preferably by some fair level of test automation.
You do need a release librarian as well as a release engineer. The release librarian will work over and above the dev and the testing librarian and will be responsible in main for labeling and maintaining the separate releases. The release engineer will need to handle the related packaging aspects. This will also ensure one more level of sanity / smoke testing before release.
The management focus will I think be mainly on the release cut-offs and the movement of source code.
Have you also streamlined the bug reporting and fixing between the dev and testing teams? Trust that will also turn out to being an icky area. (so what's new eh! [img]images/icons/wink.gif[/img]