Il 29/11/2011 14:18, RW ha scritto:
On Tue, 29 Nov 2011 10:52:58 +0100
Marcello Romani wrote:
I'm no expert (just a few squid deployments) but that doesn't make
any sense to me.
"caches the content in a normal way", i.e. on a cache miss it fetches
the object from the net and stores it for other clients, while on a
cache hit it serves the file from its local depot (I'm
oversimplifying here).
"but once a client gets a cache miss, then their Squid allows the
client to fetch the record" hmmmm....
"Once the object is fetched by the client, Squid intercepts it and
stores the object for the other clients."
How can squid "intercept" a file _after_ it's been downloaded
_directly_ from a client, if that file doesn't pass through squid
itself ?
As I understand it there are 2 servers, "Web Cache" and one or more
squid instances. WC passes the request to squid, which serves the
object or passes it back to the parent cache part of WC.
I think that one reason for this is that WC controls what goes into
each squid cache for policy reasons.
What's less clear is precisely how misses are handled. It might just be
a confusing description of :
"WC front-end proxy"<--> Squid<--> "WC parent cache"
But what it seems to be saying to me is that WC sends the origin-server
response directly back to the client, and independently sends it to
squid (with the front-end ignoring the squid response).
Much clearer now, thank you.
--
Marcello Romani
