Don’t touch systems in mid of projects.

Don’t touch systems in mid of projects.

In my last post “How many Landscapes/Regions/Systems/Environments needed for smooth deployment” I mentioned that we are changing our systems as per second solution.
However that solution is very painful. The thing we learn from that change is “Do not touch your stable systems in mid of projects.

Yes that a great lesson we are learned. We are really going through system change pain. I know if we successfully implement our changes, then system blockage problem will be over. But so far that seems to us a long way to go.

Problem we are facing with change is:
Our all projects are stuck. Why, be because all System and clients are in update mode. So we can do anything.
All QA times are consumed by smoke testing. The whole day we are just doing smoke testing in different, different systems and clients. I suggested my Project Manager, that lets first do all updates and let’s don’t make any build and let’s make all system to be little bit stable. When all dust stable then we can go and do smoke testing and other testing. But my Project Manager is no mood to listen me.
Its all confusion and disorder here. There are some 4/5 builds every day. That’s making hard to track all the changes in those build.
For every builds we are doing smoke test. I think we should stop doing smoke test for every build, until unless when if build have some major changes. Our QA team is so confused. Example, if Team are in mid-off smoke test in T0, and then come emails from PM, that now do smoke test in T1 or T3… and blabla...

Yes I know it’s painful, but I’m very optimistic, that, after all systems became stable, then life would be easy. I hope…

How many Landscapes/Regions/Systems/Environments needed for smooth deployment?

How many Landscapes/Regions/Systems/Environments needed for smooth deployment?

That was the agenda of our meeting, last Friday. Mostly this question always comes in big organization when big releases (with multiple projects) move into production. For small organization this is not a concern, but for the big companies this always a topic for discuss. For all seriousness, mostly all organization doesn’t pay lot of efforts on this very important, fundamental issue. Believe me, lot of company’s time and money can be save, if they figure out there right Landscape/Region/Systems/Environment needed for implementation.

Here, we have just one general Landscape/Region/Systems/Environment they are: T0, T1, T2, T3, TX and prod; these are alliance with SAP CML/BI Landscape with different clients (golden-100,200,300…). T0 is sandbox, T1 is for development for non SAP application and ET1 (100/300/500) for SAP. T2/ET2 (100/200) is for QA and T3/ET3 (100/300) for UAT. TX for pre prod. support and then prod. This is typical IT Landscape/Region/Systems/Environment, you will found in all big companies.

Problem:
So what is the problem with these conventional environments? Problem is when you are going for big releases, with multiple projects. Suppose you have very ugly data in T2 for SAP and you want to clean and refresh that region, then all projects (including SAP and non SAP) in that region became stand still. In these old-fashioned environments, this delay the projects schedule. Now think how big the problem will be when you have to refresh and clean data form all T0,T1,T2,T3,Tx and with different Releases and different projects, and SAP needed to transport all request. That consumes lots of time, money and efforts.

Solutions:
One solution is simple; give every project, own environments. Yes solution seem simple, it’s very expensive and very hard to maintain. It’s like impossible providing independent environments for all projects.

Second solution is, don’t give full independent environments, and instead give partial independent environment for all projects. For example, gives all projects their own partial independent environment, till QA, below UAT. In this way we can stop clogging the Environment/Landscape/System for different releases and projects and as well as refreshing and data cleaning of environment will be easier and clean.