We are migrating from WebSphere V5 to V6. Also going from Single Server mode to Network Configuration. We are going from 1 Production guest per
server to 1 Production guest for 3 production Servers. Each WAS Server i s doing the same work as before ie Server1 is App1 and so on, we are not mixing workloads. The application response time is not as good in V6 vs V5. I'm not looking for WAS help, but in the log they are receiving 'cpu starvation' messages. Linux/UNIX admins running VMSTAT say they see idle of 0. Yet when I log on to the linux guest I dont have any problems 'running' things and neither do they (linux unix admins). Also VELOCITY shows this guest averaging approx 25-50% of 1 IFL (we have 2). I'm perplexed, is this guest actually suffering from 'lack of cpu' during some peak periods? How do I verify? Any thoughts suggestions are much appreciatted.