Re: Pre-fetching some URLs

Ong Beng Hui wrote:
>
> > I have seen a lot of messages where people state that their cache never
> > fills. Whilst 30 Gib maybe a different league I had the same problem here,
> > my cache would never fill.
>
> Yes indeed.
>
> > After a bit of playing around I found if I uncommented the default
> > ttl_pattern (max ttl of 30 days) lines from squid.conf that my cache did
> > start filling up at a more reasonable rate:
> >
> > #ttl_pattern ^http:// 1440 20% 43200
> > #ttl_pattern ^ftp:// 10080 20% 43200
> > #ttl_pattern/i \.gif$ 2880 50% 43200
> > #ttl_pattern /cgi-bin/ 0 0% 43200
> >
> > I suspect that a lot of people who install squid, run it with the default
> > configuration initially to see how things work while they get a handle on
> > the operation of things. Then go back and fine tune their parameters.
>
> "ttl_pattern ^http:// 7200 80% 10080"
> "ttl_pattern ^ftp:// 10080 80% 10080"
> "ttl_pattern \.gif$ 10080 80% 10080"
> "ttl_pattern /cgi-bin/ 0 0% 10080"
>
> Here's mine which is a little less agressive.
> I am thinking of increasing the TTL to even higher,
> but I risk the possiblity of having stale objects
> in the cache.

I'm a smidgen unclear on what each of the three fields above do. The
percentage _seems_ clear enough. If the document was last modified X
time ago, then we keep it for 80% of X, if given no other expiry data.

What about the other two fields? Limiting factors? The .conf comments
are less than clear on their exact nature.

D
Received on Thu Oct 24 1996 - 20:51:05 MDT