I ran into this issue every other 4~5hours, with load of
450vusers, simulated traffic which is mixed with HTTP/HTTPS/FTP/TELNET, ~300Mbps throughput.
What is more wired is two testbeds have the same hardware and software configuration, deployed with identical silkperformer project, only one testbed occurs this memory allocate out problem.
We use two agents to spread the load, the agent spec:
win server 2003 standard edition SP2
Intel(R) Xeon(R)CPU E5420@2.50GHz