Re: Cannot cache FTP request (HTTP experts advise needed)

From: Robert Collins <robert.collins@dont-contact.us>
Date: Wed, 1 Nov 2000 08:52:01 +1100

Could we timestamp those requests when they arrive as per rfc 2616 section
14.18? (Or do we already and that code is to handle "legacy" versions of
squid?

Rob

----- Original Message -----
From: "Duane Wessels" <wessels@squid-cache.org>
To: "Dr. Michael Weller" <eowmob@exp-math.uni-essen.de>
Cc: <squid-users@ircache.net>
Sent: Wednesday, November 01, 2000 5:53 AM
Subject: Re: Cannot cache FTP request (HTTP experts advise needed)

>
>
> On Tue, 31 Oct 2000, Dr. Michael Weller wrote:
>
> > Well, yes, my actual problem is I do not understand httpCachableReply()
> > especially because it checks zillions of internal bits, whose meaning is
> > unknown to me. I also do not want to change the caching of the other
> > files. Where would I put my "return 1" in httpCachableReply()
> > to deal only with this (or only a "little" more) case(s).
> >
> > That cache peer, btw, is not a real cache but a virus scanner. It really
> > makes no sense to leave the caching to the peer here.
>
> You can try this:
>
> Index: http.c
> ===================================================================
> RCS file: /squid/squid/src/http.c,v
> retrieving revision 1.366
> diff -u -r1.366 http.c
> --- http.c 2000/07/18 06:16:41 1.366
> +++ http.c 2000/10/31 18:46:42
> @@ -255,6 +255,11 @@
> case HTTP_MOVED_PERMANENTLY:
> case HTTP_GONE:
> /*
> + * hack
> + */
> + if (PROTO_FTP == httpState->request->protocol)
> + return 1;
> + /*
> * Don't cache objects that need to be refreshed on next request,
> * unless we know how to refresh it.
> */
>
>
>
> The reason for the code commented with "don't cache objects from
> peers w/o LMT, Date, or Expires" is from some problems observed in
> a while ago in older Squid versions. Squid used to cache responses
> that didn't have any timestamps. Such an object could "live forever"
> in a sibling relationship. As soon as it got removed from one
> sibling, it would be found again in the other sibling and then get
> cached again for some default period of time. This continues
> back-and-forth and makes it really hard to purge stale responses
> from some hierarchies. So we added the requirement that if squid
> gets something from a neighbor, it must have at least one timestamp
> header so we know how long its been around.
>
> --
> To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
>
>

--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
Received on Tue Oct 31 2000 - 14:49:21 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:56:04 MST