On 28/02/18 21:30, Grey wrote: > > At this point, I'd like to know if what I'm trying to do is possible at all, > beacuse I'm starting to think there's something major I've totally > overlooked. The most obvious thing is that the port 3128 is an *HTTP* port, so when the Browser is using that port it is using HTTP protocol to a) request that Squid perform all the FTP things and deliver the result as an HTTP response, or b) create tunnels through the proxy to the relevant FTP IP:port of the origin servers. By comparison the ftp_port receives FTP protocol natively. Squid is mapping that FTP protocol messages into HTTP and handling the mapped requests. IIRC these mapping requests are only handling internally by the same Squid that created them. The outgoing is mapped back to FTP for the server contact. I used to think those requests could go to a cache_peer, but Alex indicated recently that they do not. It is technically possible to be done with a few additions to Squid. But those additions are not present in any current version. Amos _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
