I have never tried this kind of configuration, and I dont know what kind of
raid you are using, but I would not use raid for a cache filesystem...
Unless you have your raid disk created using a fast adapter and even faster
disks, IMHO, that could cause performance problems on the cache device.
Of course, I'm assuming the raid stuff due to the path for your cache
directory.
Forster
Nick Lomonte <nick@eonet.net> on 12/18/2002 07:02:33 PM
To: squid-users@squid-cache.org
cc:
Subject: [squid-users] 100% cpu
I've checked the FAQ and read over the archived messages over this (I
realize it's a common topic), but nothing has helped yet.
Running squid-2.4.STABLE6-6.7.3 on a redhat 7.3 box. Using transparent
proxying via a cisco router and iptables. As soon as it starts caching,
the CPU jumps to 100% and stays. It doesnt do it every 5 minutes etc,
it's just constant. The performance SEEMS to be ok with it, but it
still worries me for obvious reasons. Below are the only changes I've
made to squid.conf:
http_port 3128
httpd_accel_host virtual
httpd_accel_port 80
httpd_accel_with_proxy on
httpd_accel_uses_host_header on
cache_replacement_policy heap LFUDA
memory_replacement_policy heap GDSF
cache_dir aufs /raid/cache 100000 30 256
cache_access_log none
cache_store_log none
redirect_children 10
Any ideas?
-- Nick Lomonte Network Engineer Eonet nick@eonet.net 409.833.1700Received on Wed Dec 18 2002 - 14:10:08 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:12:08 MST