httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
Subject Re: [users@httpd] Problems with ProxyPassReverseCookieDomain
Date Mon, 15 Oct 2007 13:06:05 GMT
Distinguishing between FQDN and Domain is barely relevant when
discussing Cookies.  RFC2965 states, "Domain Defaults to the
request-host."  Cookies default to using the FQDN as the Domain if the
Domain is not specified.  Specified domains must start with a period,
must be exactly one level below the server name, and are only returned
to servers in the one level above the period.  Read Section 3.3.2 and
weep.  The drafters deliberately made Cookies as limited as possible. and cannot share Cookies even
though both addresses reach the same server.   Look at the Cookies in
your browser.  The server name will be part of the domain if no domain
was specified (e. g. and

I wonder if the leading period is missing from "".
"" (leading period) would be sent to and  "" (no leading period) would only be
sent to  Browsers should add the leading period if domain
is specified.


On 10/15/07, Axel-Stephane  SMORGRAV
<> wrote:
> The behaviour observed by the original poster is correct: is the host
FQDN - not the domain. The cookie _domain_ is
> A cookie which domain is will be submitted by the UA to both
and The conclusion is that no rewriting of the cookie domain is required.
> However, the original poster states "it is named simply" which makes me wonder
if he/she confuses the cookie name with its domain. In order to clear that up, it would be
nice to see a trace of the requests/responses between the UA and the reverse proxy.
> -ascs
> -----Message d'origine-----
> De : [] De la part de
> Envoyé : lundi 15 octobre 2007 10:11
> À :
> Objet : [users@httpd] RE: Problems with ProxyPassReverseCookieDomain
> An "" Cookie will be sent to and
> Questions for troubleshooting:
> Is receiving the Cookie?  (Should be yes.)
> Is sending the Cookie when proxying to
>  Is receiving the Cookie?  (Both
> should be yes, but mod_proxy breaks in every other release.  This
> 2.2.1 changes state this issue was fixed.)
> Does the server software on check the Domain field on
> incoming Cookies and ignore Cookies from parent domains?  (Would mean
> overengineering. If the Cookie arrives, use it.)
> When you write "parameters", do you mean HTTP Headers, QueryString
> parameters, POST parameters, part of the URL, or something else?
> The parameters issue may require mod_rewrite.  Ttry this:
> RewriteEngine On
> RewriteRule ^/site/(.*)$$1 [P]
> solprovider
> ===
> From: Skiff Lad
> On: 20071002 14:59
> I seem to have a problem / misunderstanding regarding the use of the
> ProxyPassReverseCookieDomain directive on 2.2.4 on Windows.
> I have running 2.2.4. Within the httpd.conf file, I
> have a number of redirect directive statements that send a user, based
> on the URL, eg. to etc. This
> all works fine so that the user arrives on the www1 box and there is a
> cookie correctly set with name
> It now transpires that I need to use ProxyPass to pass the parameters
> coming in on the request to onto - as
> redirect doesn't send these initial parameters across.
> So, my httpd.conf file (on now looks like:
> ProxyPass /site1
> ProxyPassReverse /site1
> ProxyPassReverseCookieDomain
> But nothing seems to be happening. When I access
>, there is a cookie correctly created (with the
> right contents etc), however, it is named simply, ie.
> there is no www. nor www1. prefixing - this breaks subsequent
> processing on the box because there is no cookie
> named accordingly.
> What am I doing wrong ? Do I need anything else ? Or am I being plain stupid ?
> Thanks
> Skiff

The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:> for more info.
To unsubscribe, e-mail:
   "   from the digest:
For additional commands, e-mail:

View raw message