Tuesday, June 22, 2010

Test, QA and Production SharePoint infra should be comparable

In multiple SharePoint projects I have been confronted with delivered functionality succesfully validated in the test environment, does not operate the same in the QA environment. Since the application software is identical in the environments, the cause of the malfunctions must originate from differences at infrastructure level. Especially when it occurs in a later project stadium, this can be a very unpleasant surprise. Of course it's not realistic to expect nor demand that the Test, QA and Production environments are 100% comparable. However, the differences should be limited to hardware resources level; number of front-end servers, sizing of database, cpu, ... The infra architecture and configuration should be [logically] the same; load-balancing, firewall settings, proxy, Central Admin settings (e.g. Search, SSO, Policy for Web Application).
Extra complication to achieve such an ideal Test/QA/Production setup, is that typically the Test environment is much more volatile. This is the first environment to try-out your stuff, be it application functionality, infrastructure settings, or both. And in case of failure, not always the deployed trial stuff is rolled back. This is an extra argument to use virtualized Test environments per SharePoint project. When a new project start, an own Test environment is provisioned, with the current production state as its baseline. The project deploys all the project deliverables to this own Test environment. There is no overlap or conflicts with the work from other projects. And after project end (successful completion, or earlier stopped), the project Test environment can be deactivated and retired.

No comments:

Post a Comment