Community Tip - New to the community? Learn how to post a question and get help from PTC and industry experts! X
Hello Patrick.
I will not pretend what we are doing in my company is best practice but this is what we had to consider.
First, back in 2007. we had a blank environment as we never had Windchill. We bought physical server (Virtualization was out of reach at that time). We first started with a sandbox install on a server. The implementation team consisted of people from different location.
Another thing I got straight from the beginning is NO CUSTOMIZATION. Windchill is so far used mostly to store CAD data and Document at the moment. Windchill is configurable enough to use different type of numbering scheme that customisation is not required or business should adapt.
So we had a sandbox. then we moved to what we call Test Production Server (TPS) with replica server and much powerfull server.
Finally, we went live on the same type of servers. What could be exported from TPS to live server was exported. Other things needed to be done again. Which quality issues of course. Doing the job twice is never a good thing. Getting it done right first time is no guarantee of success the second time. PTC could develop some sort of transport as you regularly find for ERP.
One year after go live with 9.0 we moved to 9.1 and also IT had purchased VMW. So we moved TPS and production to this environment.
However, VMW for us meant sharing. IT got a big box with so much resources in it (Memory, Disk space, CPU) and thanks to VMW technology it is very flexible, one can add CPU, Memory to an existing server very quickly, servers can be created also very quickly. After 3 years of use and as our Windchill environment grown, we found ourself to have to "fight" more and more for resources.
We are no moving to 10.1. We upgraded our sandbox from 9.1 to 10.1 on the same virtual sandbox, but the TPS and live will have dedicated physical server. Yet we can still set up a virtual environment on those physical server. The decision has not yet been taken. but the bottom line is that I now have dedicated black boxes high specs just for Windchill.
As you can see we have always used servers for everything we do from development, testing, User Acceptance Testing. Mainly because the implementation is located in different location and work at different time.
Also we are working a lot with our Consultant. Therefore when we ask our consultant to develop a script to let's say change in mass the life cycle state. I do not know how he works. Maybe he has his own Windchill on his laptop that he keeps reinstalling everytime he crashes it. What I know he that he will develiver the script on our sandbox or TPS (depending on the risk and level of testing required), We will test it and once we are happy, we install it on the production.
I hope this give you some useful information.
Best regards
Our company (GPSL) does Windchill customization services. Our practice is to develop locally by including the codebase jars in the classpath. This provides full access to the API. The same goes for Info Engine.
For testing, we have a pool of Windchill servers that we deploy the code on for testing.
We recommend our customers have a beta environment where they can deploy and test new code and configurations. If they don't have such an environment, we replicate their production on one of our servers.
I used to keep a blank install of Windchill local, but that got too cumbersome as we were working with clients on different versions. By using servers we can support multiple versions at one time. This especially comes in handy when regression testing new versions.
Servers also provide better hardware specs. Good performance is nice even when testing.