On Mon, 25 Oct 2004, Laurent Evrard wrote:
the larger the blocking file, the higher the CPU usage: Each request to Squid has to parse this file, resulting in high CPU usage;
Only if you have made an extremly poor choice of acl type.. i.e. using url_regex where you should have been using dstdomain and other more structured acls.
I suggest you employ generic words instead of maintaining hundreds of thousands of sites!! Using generic words when possible shall result in lower CPU usage.
Or the opposite if this word list requires the use of regex based acls where one could do with non-regex acls when specifying the complete list.
Regards Henrik
