Hello, I am trying to squeeze a little more performance out of the squid in http-accelerator mode and wondered about how the persistent connections worked? When I first set up the server. I had server_persistent_connections on But then I noticed that each apache connection would only serve a single request then stayed open till the apache keep-alive time-out. My process list was huge. When I turned off persistent connection in apache and squid, the process list went down dramatically and the websites were much more responsive. I envisioned that squid would open up a connection and funnel requests through it serially until one side closed the connection. Is this not the case? I have looked at directives like pipeline_prefetch but this doe not seem like it will have the desired effect. Any help would be appreciated. Thanks, Jerry -- Jerry Thomas thomasj@xxxxxxxxxxxxxx I.T. Director ph. 305-662-5959 ext. 242 The SCORE Group fax 305-662-8922
