mam serwerek ze squidem dla kilku komputerow i =
znaczaco obciaza mi procesor
robilem testy juz na dwoch maszynkach
1. 2x p3 700 (1,5G RAM SDRAM , 2 HDD SCSI 40G )w =
80%
2. AMD XP 1800 (256MB DDR, 1HDD IDE ) w 50 %
konfiguracja squida w obu wypadkach taka sama
100MB w ramie
i 5GB na dyskach
czytalem za squid gdzies obslugiwal 400 polaczen na =
sekunde, a u mnie przy
40/sek juz takie obciazenia
moze cos skopalem w konfiguracji ?
sample_start_time = 1064610218.763519 (Fri, 26 Sep =
2003 21:03:38 GMT)
sample_end_time = 1064613829.986447 (Fri, 26 Sep =
2003 22:03:49 GMT)
client_http.requests = 40.010546/sec
client_http.hits = 20.294510/sec
...
client_http.kbytes_in = 17.286665/sec
client_http.kbytes_out = 174.967874/sec
...
server.all.requests = 20.464259/sec
server.all.errors = 0.000000/sec
server.all.kbytes_in = 147.499894/sec
server.all.kbytes_out = 13.147624/sec
server.http.requests = 20.462597/sec
server.http.errors = 0.000000/sec
server.http.kbytes_in = 147.494633/sec
server.http.kbytes_out = 13.146239/sec
...
unlink.requests = 6.105965/sec
page_faults = 3.245438/sec
select_loops = 141.947758/sec
select_fds = 266.759217/sec
average_select_fd_period = 0.003699/fd
...
swap.outs = 6.962461/sec
swap.ins = 22.900276/sec
swap.files_cleaned = 0.549675/sec
aborted_requests = 4.406817/sec
syscalls.polls = 159.981815/sec
syscalls.disk.opens = 18.413984/sec
syscalls.disk.closes = 18.412599/sec
syscalls.disk.reads = 17.044088/sec
syscalls.disk.writes = 42.680002/sec
syscalls.disk.seeks = 0.612535/sec
syscalls.disk.unlinks = 0.000000/sec
syscalls.sock.accepts = 34.017008/sec
syscalls.sock.sockets = 28.072485/sec
syscalls.sock.connects = 22.596500/sec
syscalls.sock.binds = 22.612838/sec
syscalls.sock.closes = 35.706464/sec
syscalls.sock.reads = 155.345436/sec
syscalls.sock.writes = 114.475071/sec
syscalls.sock.recvfroms = 2.792683/sec
syscalls.sock.sendtos = 1.477893/sec
cpu_time = 1844.880000 seconds
wall_time = 3611.222928 seconds
cpu_usage = 51.087403%
tam gdzie byly zera wyciolem linie