SolMan and CMMI Configuration Management

Question

Hi
We are looking into implementing CMMI level 3 in our organisation.
One of the main points is to use SolMan for Configuration Management which seems obvious.
Does any of you "out there" have any experiences and answers/ideas to the following topics:
- which CI's have you identified?
- how are versions of the CI's managed?
- how are you managing baselines and is there any demand for a baseline "rollback" in CMMI?
And how are you handling Requirements Management and Requirements Development in SAP Solution Manager? Especially the bidirectional traceability
Have you found any litterature on the subject?
Thanks in advance - all help is welcome
Peter

Answer 1

I don't have a good answer to your question, but chose to reply because I'm also in search of Architectural and Operational standards and practices for WLS. Everything I see in terms of Oracle and 3rd party documentation, including the product itself, seems geared toward the developers and optimized for for their needs and desires. Precious little advice in terms of best practices for deploying and hosting production services on this platform. I'm just a lowly SA and need things explained like I'm retarded.
Obviously common sense will have to end up addressing most of this, but why must we re-invent the wheel?
FWIW in regards to your question, we don't rebuild our environments from scratch at every deployment. We don't even delete configurations and components at deployment. (Although I've heard that's a recommended practice, I never get a straight answer for "why?" From this I'm left to conclude that; what, WLS is unreliable, unstable? Until I hear a good reason why I should be deleting everything all the time I'll stick with my instincts.) As for configuration management, we're stuck in the stone age I'm afraid; although high on my list is to start reverse-engineering our prod config into scripted form so we can do just what you're explaining.
At deployment, for every object we check for it, add it if it's missing, and re-set the configs. For all application deployments we undeploy/redeploy. All in WLST scripts that live in our version repo, natch. We could be doing it totally wrong, but it works so there's little impetus to change it. We're also handling an OSB deployment in a similar fashion, but that's done with a heavily modified-for-portability import.py script w/ ant.