For some reason, the concept of conducting full function tests prior to ending the Lync POC or pilot project has come up again and again. Those pesky customers just keep insisting.
Usually, the customer has already come up with a fairly exhaustive test plan on their own and all I need to do is help them revise or add expectations. If they have not developed their own test plan, I first point them to the Lync 2013 RASK.
The Lync Rollout and Adoption Success Kit can be found here. If you poke around the semi-convoluted structure you will eventually divine the logic, but you are probably better at that than I am. At any rate, eventually you will
find stumble upon this link: http://www.microsoft.com/en-us/download/confirmation.aspx?id=37031 which is the download page for the Lync 2013 Rollout and Adoption Success Kit (RASK) Resources package. A very nice piece of kit. This download has the following format:
I have taken the liberty of showing the location of the subject for this article – the “Sample Pilot Test Plan.xlsx”. What you see below is a modification of that fine piece of work. In its’ base form, the RASK test plan has saved me many hours of skull sweat – and most likely saved my customers many hours as well. For a simple Lync deployment, you may be able to use it as is.
But the real value of the RASK test plan is to get your head into the game. Accordingly, my modifications are just that, MINE - as needed for a project, and then modified past that to serve as my personal baseline test plan.
If you were to run a full system test and that included HA and DR, you would also iteratively fail Front End Pool members and run full-tests per this matrix with each pool member offline in turn. And then you would need to create the same set of tests with perhaps one Edge Server offline and then consider your other environment details. And then loop through for each pool member off in turn et cetera. You may end up with a matrix with a considerable number of test cells to complete. I one time did a project where the core HA/DR test matrix resulted in 900+ individual test cells. Each cell was one complete test that contributed to the overall upper level test. Fun! But in the end, if someone asks you, you can show them, yes, we tested.
I try very hard to accomplish the testing inside out and and using this order of tests:
- Test the FE pool level - this may include monitoring, archiving, persistent chat, and Office Web Apps Servers
- Test the Edge (Edge servers and reverse proxy)
- Test mobility clients
- MPOP scenarios
Resist the urge to use real users. Gin up a set of test accounts, per pool, and use those accounts for the testing. If you MUST use real accounts, be in control of the testing flow or you will be like this when things don’t go exactly right and the testers start veering off into the own little interest area.
So, without further ado, here is a sample what I use as my baseline in screen cap format – a download link can be found at the bottom of this article.