Re: peering

From: Dancer <dancer@dont-contact.us>
Date: Tue, 05 May 1998 15:19:17 +1000

Antony wrote:
>
> Hi
>
> We've set up a new peering link where the peer is allowed to access our
> cache as a sibling for cache hits, but is not allowed to pull through out
> cache.
>
> I am seeing this...
>
> 894341279.884 1393 203.22.101.5 TCP_REFRESH_HIT/200 619 GET
> http://www.colorado.edu/images/nav-info.gif -
> PARENT_HIT/proxy2.syd.connect.com.au image/gif
>
> Which to me looks like they are able to pull through our cache (thus costing
> us money). Am I right or wrong on this point?
>
> We have other peers, everyone is configured exactly the same, yet this new
> link is the only one showing up like this.
>
> Any advice would be appreciated. We are running squid-1.1.20

ICP will give you a certain (small) percentage of false hits. These
things tend to go both ways, and the little bit that you fetch through
each-other from those false hits is nothing to worry about. That
percentage is drastically increased, however, if you and your peer use
different cache policies (refresh_patterns), then you might wind up with
a worrisome imbalance.

D
Received on Mon May 04 1998 - 22:32:07 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:40:05 MST