I'm pretty new to Silkperformer, but I was wondering if someone could possibly provide me with some information. I'm running a test with 70 users with WebSphere 5.0/SQL 2000. It varies in the tests, but every so often, I get users dropping out with "HTTP: 1011 - unexpected connection close during read., SendRequest second attempt." It varies between the user and url. I'm unable to find any information on this error or how to get around it. Any Ideas????
I have come across something similar to this and found the problem to be the actual office LAN network being maxed out... Finally mangaged to solve it by using 2 external agents routed through a 4 gb switch. Maybe your problem could be similar?
I also see this when an agent machine is trying to simulate more users than the NIC card can handle. The wizard that estimates how many users you can simulate on a machine only takes memory and CPU into account. If your network card can't handle the load though, timeouts will occur.
We've seen the similar problem durning our testing too. We found out 2 reasons for that Number 1 is already define in previous posts, which is NIC related. The second one is application server itself, If the applcation server hit the bottle neck any can't send the responce quick enough then this condition also happen.