[squid-users] TCP connection to 127.0.0.1/80 failed

Hi,
Some error occured when I am using http_load to test my squid server

----
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| TCP connection to 127.0.0.1/80 failed
2008/07/03 18:07:42| Detected REVIVED Parent: squid_test_ip_6
2008/07/03 18:07:46| squidaio_queue_request: WARNING - Queue congestion
2008/07/03 18:07:54| squidaio_queue_request: WARNING - Queue congestion
2008/07/03 18:08:07| squidaio_queue_request: WARNING - Queue congestion
2008/07/03 18:08:29| squidaio_queue_request: WARNING - Queue congestion
2008/07/03 18:09:13| squidaio_queue_request: WARNING - Queue congestion
2008/07/03 18:10:42| squidaio_queue_request: WARNING - Queue congestion
I am sure back-end server(127.0.0.1/80) is running and have no problem
when this happen
Maybe http_load make too much pressure that squid can't afford.But I
hope squid can afford the pressure by use the max cpu, memory or any
other thing of the computer
How to do the job and avoid the error happen?
Thanks