On Fri, 7 Nov 1997, Jtn wrote:
> We are a small ISP pulling in about 600MB per day and would like to use Squid to reduce this.
>
> I've set up Squid on our web server:
> It has a 4GB SCSI devoted to the cache and 96MB RAM (Although for some reason Linux 2.0.30 'free' only reports 64MB even though the CMOS reports the full 96MB)
in lilo.conf, in the section for your boot image, put
append="mem=98304K"
which will tell Linux to up the memory to 96Meg (64meg is the max
returned by BIOS) once you rerun lilo and reboot.
FreeBSD ppls need to put:
options "MAXMEM=98304"
in /usr/src/sys/i386/conf/whatever and recompile.
(perhaps this needs to be in the squid FAQ ? )
> I've played around with mem and swap sizes and can't seem to get the cache to grow past about 13MB.
> VM-Max 15 Maximum hot-vm cache (MB)
> Swap-Max 3000 Maximum disk cache (MB)
> cache_mem 15
> cache_swap 3000
Increase cache_mem perhaps? ;)
--==--
Bruce.
Whose currently scratching head over NUMBER_FDs > MAX_FDs despite
increasing the linux kernel and decreasing squid's.
Received on Thu Nov 06 1997 - 17:39:54 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:37:27 MST