Re: why send ICP queries for hierarchy_stoplist matches?

Don Lewis writes:
> I'm running Squid 2.2.STABLE4 behind a firewall, so I have it
> configured with a never_direct acl that matches anything outside the
> firewall in order to get Squid to fetch through its parent cache, which
> is outside the firewall. I noticed that Squid does an ICP query for
> URLs that match its hierarchy_stoplist, which it would normally fetch
> directly if it weren't behind the firewall. This ICP query seems kind
> of wasteful, since the parent cache won't have the object cached. It
> seems to me that it would be better to just do the http query for
> anything that is known to be uncacheable.

I believe the issue would be that without knowing the parent's acls, it
doesn't know whether the parent would regard it as uncacheable. There
are also cases where even with never_direct, it should still do ICP
queries, because it might need to find which of several parents or
peers might already have it.

The real optimization question is a little more subtle: if it's certain
that regardless of ICP query results, Squid will end up going to a
single proxy to fetch a URL, should Squid skip the ICP query and cut to
the chase? It's a "special case", but probably one of the more common
ones.
  -- Clifton

-- 
 Clifton Royston  --  LavaNet Systems Architect --  cliftonr@lava.net
        "An absolute monarch would be absolutely wise and good.  
           But no man is strong enough to have no interest.  
             Therefore the best king would be Pure Chance.  
              It is Pure Chance that rules the Universe; 
          therefore, and only therefore, life is good." - AC