On Fri, Aug 06, 2010 at 05:43:03PM +0200, zabrane Mikael wrote: > > As I wrote it before, accessing our backend directly with "curl" > or any web browser worked for us for years now. Here's an example: > > $ curl -0 -D - -v -I --no-keepalive --proxy 127.0.0.1:7676 > http://www.groupama.fr/ > Date Tue, 15 Jun 2010 09:18:29 GMT > Content-Type text/html; charset=UTF-8 > X-Powered-By ASP.NET > Set-Cookie > jsessionid=MXFJmqLZ2cwnR1gyFhVDvzG59zx2KcP2vKcg6lNP4Gnz5nVky4D1!1556069365; > path=/ > [...] > >> It seems like Varnish is only adding the erroneous header when it is >> using your "caching backend", yes? (i.e. the thing on port 7676) > > Yep, only with our backend. > >> Would capturing the traffic Varnish sees with tcpdump and comparing >> the two (squid and the caching backend) be feasible? > > with wireshark, it's my first time using it (please, see enclosed files).
Sorry, I was more interested in the actual data of the response. Presumably Varnish is interpreting the response from squid differently to the response from the caching backend. Or, is there a way for us to see the response from the backend? Is the site http://www.groupama.fr/ currently served by the caching backend, i.e. if I point a Varnish at it I should see the same bug? _______________________________________________ varnish-misc mailing list [email protected] http://lists.varnish-cache.org/mailman/listinfo/varnish-misc
