I'm running squid squid-2.4.STABLE7 on a sparc machine (4x18 scsi HDD)
with Openbsd 3.1. I have 2000 users using it, my Alteon switch is
configured for 600 concurrent connections. After 2 days, users begin
complaining about slowness, so I have to bypass the cache.
"top" shows on a high 53% utilization and 0% swapping during peak hours.
I also increased my file descriptors to 16384.
My squid.conf:
cache_mem 128 MB
cache_swap_low 90
cache_swap_high 95
maximum_object_size 2048 KB
cache_dir ufs /cache1/cache 6000 14 256
cache_dir ufs /cache2/cache 6000 14 256
cache_dir ufs /cache3/cache 6000 14 256
half_closed_clients off
memory_pools off
# ulimit -aH
time(cpu-seconds) unlimited
file(blocks) unlimited
coredump(blocks) unlimited
data(kbytes) 1048576
stack(kbytes) 8192
lockedmem(kbytes) 564880
memory(kbytes) 564880
nofiles(descriptors) 16384
processes 1044
Besides this all the rest is default settings.
Am I pushing the server limit or it can handle all those users?
anything else that I should look for?
Thank you.
Received on Thu Dec 05 2002 - 07:59:33 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:11:52 MST