AIX

Hi Guys,

I need some help analyzing the attached metrics. System context is 2 LPAR's in a P795 running WebSphere App Server across 4 nodes (2 on each LPAR).

Over the weekend both LPAR's lost power and upon re-start the application server response times have degraded by 25-30% for no obvious reasons.

Our middleware guys tell me that the LPAR's recovered OK and are running fine. However from an application response time perspective this is clearly not the case.

I have attached 2 text files capturing a bunch of metrics during the peak period and would be really interested in your expert thoughts.

Thanks in advance.

From the logs I do see that memory bottleneck is high , as you LPARS lost power and came back ,so
- Was the configuration same as it was earlier ? ( I suspect mem/cpu could have changed after shutdown/activation..as it wasnt rebooted )
- What does your errpr log say , do they report any error
- CPU reports average usage of around 60-70% , but did you had a look on what exactly taking those CPU cycles , what does your topas show