[squid-users] Proper Cache Peer Access

[creditu at eml.cc]

On Tue, Dec 13, 2016, at 06:33 PM, Amos Jeffries wrote:
> On 14/12/2016 2:10 p.m., creditu wrote:
> > Looking for the best way to provide cache peer access for two urls on a
> > 3.1 accelerator.  For example if a set of backend servers fullfill
> > requests for both www.example.com and www-legacy.example.com is the
> > following the correct way to handle them in regards to the cach peer
> > access? 
> > 
> 
> What you have works and is fine for simple setups like yours.
> 
> However, since you ask for "proper" ...
> 
> <snip>
> > 
> > # Backend servers for www and www-legacy
> > acl www dstdomain www.example.com
> > acl www-legacy dstdomain www-legacy.example.com
> 
> Recommended practice when you have same-type data and ACLs used in
> identical ways like these ones. Is to place both those domain values in
> the one ACL named 'www'. That will simplify your access lines.
> 
> There are some tiny memory and (cumulative) speed gains. But the biggest
> reason is easier understanding and maintenance of the config if/when it
> gets more complex.
> 
> Amos
> 
> _______________

Ah, so you would do something like this:

acl www dstdomain www.example.com www-legacy.example.com
cache_peer 10.10.10.1 parent 80 0 no-query no-digest originserver
round-robin
cache_peer_access 10.10.10.1 allow www
cache_peer_access 10.10.10.1 deny all
. . .

I was worried about a situation that was detailed in
http://wiki.squid-cache.org/SquidFaq/SquidAcl#And.2FOr_logic (the src
address example).  Thanks