Re: Approach/Order of testing
Wow! that is a lot to do. If you have the resources, I would try to set up two test environments. One for regression testing and one for load/performance testing.
Alternately, you could set up some skilled testers/users to begin regression testing while another team works out a plan for and prepares the performance and load testing.
And if things go wrong on the performance/load end, you can still test for basic functionality and bug fixes while the those issues are being smoothed out.
Approach/Order of testing
I have been asked to devise a testing approach and could also use some help on the order in which testing should occur. Here's the situation. We currently have a vendor supplied client/server app in production today that has maxed out the number of cases and users, so has to be re-written. This 1.0 version holds info in memory and some in a DB2 database. In addition to this app, we have in-house developed pieces - a middleware which makes MQ/CIM calls to our mainframe for addl data, a web inquiry tool and various batch jobs.
So the vendor app is being re-written to now execlusively use the database for real-time processing. The functionality will be the same however tons of bugs have been fixed. We can also now run 3 instances of the service on each server in order to improve scalability. So performance/volume testing is important. Our envionment is going from NT to 2003, plus from a clustered server hosting all pieces to 3 application servers (with a load balancing router), a DB server, a MQ server and a file server. All of this hardware is new, plus I'll be getting setup in UAT, Prod and DR enironments. So making sure it's all setup correctly is important. Plus the desktops are going from 2000 to XP.
Any thoughts in how to approach all that needs to be tested and in what order? I want to make sure we test smart since there is so much to do. Thanks.