Re: [squid-users] Problem to login to IMP WebMail through the Squid

       Hello Graeme,

Graeme.Wood@ed.ac.uk ha escrito:

> Hi,
>
> I don't think this has anything to do with squid. You need to contact
> your webmail administrators and get them to fix their IMP installation. It
> sounds like they are running two separate horde databases and your cookies
> are not being properly registered so that either system will honour the
> cookies of the other.
>

I don't think it's a IMP installation problem, because we connect to and use
the WebMail without problem when we connect to through others ISPs, or
even if we connect from our DMZ without SQUID. The problem occurs only
with SQUID.

    I don't know the IMP horde, but it seems this is a scalable and supported
architecture in order to get more throughput. Really, the WebMail admin says
me that now the WebMail server get involved three machines: one for proccesing
the mail and the two others -configured at 'round robin'- as WebMail interfaces.

> On Sat, 5 Oct 2002, Sebastian Viudez Ortega wrote:
>
> >
> > We have a problem when we attempt to read the
> > mail from our internal net to a external imp webmail
> > -we get out from our internal net through the Squid
> > 22STABLE4 at our DMZ-.
> >
> > The problem is that sometimes we don't get to login
> > to the webmail and the times we get to login, we can't read
> > the mail. When we attempt to read a message, the webmail
> > logout us with the message: 'your session has expired. Please,
> > login again'.
> >
> > This worked 2 or 3 weeks ago, and we have
> > nothing changed at our SQUID configuration. We have
> > contacted with the external WebMail Admin and he say us
> > that they have changed recently their WebMail architecture:
> > now they have two webmail interfaces in two machines
> > configurated as 'Round Robin' at the DNS. It seems a
> > related 'cookies' problem.
> >
> > Afther this, we have directed our attention to the Squid
> > and we have disabled the cache at all, with:
> >
> > acl all 0/0
> > no_cache deny all
> >
> > but the problem continue.
> >
> > ¿Does somebody know any similar problem? ¿Does somebody know
> > any trail to solve it?
> >
> > Thanks!

        A lot of thanks anyway. S. Viúdez.

--