We have online application where most of the screen programs read/write to a REMOTE tsq.We thought that making the remote tsq local would make an improvement in the cpu/response time and therefore the change the implemented in prod.However in prod env,we started hearing that the other tasks had a very very bad response time and the system guys said that the local tsq was actually ENQing resources and not letting the other tasks go thru fast.
This is hard to understand asto how is this tsq causing issues with response time when there are so many other local tsqs being processed.We backed out the changes and things started looking good.
We are trying to recreate this issue in our TEST env but have not been successful.Any ideas what the technical reason could be and how to best approach this issue to find the root cause.
Joined: 23 Nov 2006 Posts: 19270 Location: Inside the Matrix
Is it possible that your test environment is not going to be in a situation where there might be an enqueue conflict (lack of volume, lack of variety)? I suspect that the "new" process runs pretty much the same in both enviaonments, but is "blocking" other processes in the production environment.
It may be that your "new" enqueue is locking a common resource that impacts other shorter-running processes.