Suggestion required on a Contrast Load test result
I am currently involved in load testing a Java Application two boxes using Load runner.A old box and a New box.We are planning to replace the Old low Config(Hardware)box by the New High config(Hardware)box.During load testing for a load of 60 users the resposne time,hits per second and through put of the New server is very poor when compared to that of the Old server.Here i would like to add few answers to the common FAQs raised.
*Both the servers are located within the same Zone.Same location.Same room.
*Both the servers are pointing to the same database.
*Both the servers are using the same users in the parameter list
*The Vuser script is same except the pointing URLs(Old and New box URLs)
*The codes used in both the boxes are same.
*The OS in both the boxes are Sun solaris.
*The Old box is with low hardware config and New Box is with higher hardware config.
I did analysis on the analysis files exerted after the load test.I drilled down the "down load time" and found the "receiving time" of the both the servers.The receiving time taken by the New box is more when compared to the Old box.
I tried to find the actual reason for the increase in "receiving time" by deploying few Sitescope monitors CPU Utilization,Diskspace,Memory,Network,Unix resorces,Processes etc.But I could not figure out any drastic differences in the stats from the reports generated other than the "Processes" running in the boxes.
The only positive aspect about the New server box is.. This New box is capable of handling more load when compared to the Old server box.
I am not sure about the reason for poor performance of the New High config box.I need suggestions to find the exact reason for the slump in this New box.
Kindly let me know your suggestions and thoughts on this.
Thanks in Advance.
Re: Suggestion required on a Contrast Load test result
Please do locate the LoadRunner forum.