I am using 2.5STABLE3 I have a same problem over here. I have set half_closed_clients to off in squid.conf
client_http.requests = 210.684248/sec client_http.hits = 126.796477/sec client_http.errors = 0.000000/sec client_http.kbytes_in = 135.014952/sec client_http.kbytes_out = 1410.308364/sec client_http.all_median_svc_time = 0.092188 seconds client_http.miss_median_svc_time = 1.542425 seconds client_http.nm_median_svc_time = 0.012346 seconds client_http.nh_median_svc_time = 1.628034 seconds client_http.hit_median_svc_time = 0.024508 seconds <snip> select_loops = 64.277153/sec select_fds = 1312.027761/sec average_select_fd_period = 0.000756/fd median_select_fds = 16.933594 <snip> cpu_time = 281.290000 seconds wall_time = 300.055655 seconds cpu_usage = 93.745942% I have less than 5 acls's in my configuration. Venkatesh K ----- Original Message ----- From: "Henrik Nordstrom" <[EMAIL PROTECTED]> To: "Giulio Cervera" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Saturday, December 20, 2003 5:34 AM Subject: Re: [squid-users] Strange High CPU usage > On Fri, 19 Dec 2003, Giulio Cervera wrote: > > > select_loops = 280.262863/sec > > select_fds = 1502.051748/sec > > average_select_fd_period = 0.000660/fd > > median_select_fds = 3.984375 > > And you are positively sure half_closed_clients is off? > (I do remember you saying it was, but just want to make sure) > > The above makes me suspect there is some filedescriptors getting stuck as > buzy in the poll loop, which can on some systems be caused with > "half_closed_clients on" (the default). > > REgards > Henrik > >