On Tue, 18 Nov 2003, Philip Choy wrote: > Expected # of users of these cache: 1000, and currently 2 pilot users. > > The specification that I have :- > 1xE1 2mbps connection > 2GB Physical RAM > 2GHz Xeon > RHL 9.0 > squid 2.5.1
Ok, but if you are installing a new Squid why are you not using the current STABLE version? > /var/spool/squid mounted as default ext3 on 1 logical vol of phy vol group > of 4 RAID-0 SCSI disks of 20GB. Ok, but it is easier to adminstrate if you mount each drive spindle separately, and refer to them by separate cache_dir lines. > cache_mem 768 MB Why this very large amount of cache_mem? > cache_dir ufs /var/spool/squid 32000 64 256 Your really want to use one of the async cache_dir types if you are looking for any kind of performance (aufs or diskd, aufs generally preferred on Linux) Regards Henrik
