On 10 Sep 2008, at 14:19, James R. Leu wrote:
Are you sure it not actually a DNS issue? Check the URLs being sent
back
by the cat app. Do they contain the FQDN you are expecting? Are you
able to
reproduce this issue on multiple IE machines? Have you tried running
wireshark on the IE machine to watch exactly what it is doing on the
wire?
Good luck
On Wed, Sep 10, 2008 at 05:51:39AM -0700, snookums wrote:
No authentication is being used right now. I've also checked my
apache config
and used the LiveHeader add-on to Firefox. The keep-alive headers
seem to be
just as they should be. I'm beginning to suspect that the long
processing
time before the application can return content is the culprit. It
can take a
while before the body is returned. The browser consistently tries
for about
10 seconds and then throws a dns error screen up. I tried placing
$c->engine->write calls throughout to keep the connection active,
but it
doesn't seem to have done anything. Thanks for your reply.
Dami Laurent (PJ) wrote:
-----Message d'origine-----
De : snookums [mailto:[EMAIL PROTECTED]
Envoyé : mercredi, 10. septembre 2008 12:03
À : [email protected]
Objet : [Catalyst] IE has problems with catalyst
I hope someone can point out to me what's wrong here. I have a
catalyst
application that I've been developing. It takes a submitted
link, processes
it, and regurgitates it with some changes. It's a get request. The
processing can take up to about 10 seconds. I thought
everything was going
fine, but I'm developing on mac. I've been testing it in
Safari. I quickly
found that IE users couldn't access my application. They can get
to the
front page, but when they submit a link, the browser runs for
a little while
and then gives a dns error page: cannot access this page, page not
available, etc. Apache error and access logs don't even show the
hit.
I looked online and found that the problem might have
something to do with
the keep-alive option. I changed my main app file to look like
this.
use parent qw/Catalyst/;
use Catalyst qw/-k
-Debug
ConfigLoader
Static::Simple
UserAgent
Unicode
Compress::Gzip/;
our $VERSION = '0.01';
That didn't fix the problem, but I noticed some improvement for
slower
connections that weren't even getting the front page before. I
also should
mention that I'm using Catalyst with the latest versions of
perl, apache,
and mod_perl. I tried changing a few of my keep-alive options
in apache conf
as well to increase the time before dropping a connection. I
haven't noticed
any difference.
Can anyone tell me how -k works? Is there a default timeout
value that I can
change? Is this even the problem?
--
Hi,
If your hypothesis about needing Keepalive is correct, and if your
application in production runs under Apache, then you you need to
activate
Keepalive in your Apache configuration; its not a Catalyst
property. So
edit your httpd.conf and add
KeepAlive On
One reason why this might be important is if you use the Microsoft
NTLM
protocol for
authentifying your users (mod_auth_ntlm or mod_auth_sspi or
PerlAuthenHandler Apache2::AuthenNTLM); these definitely need
KeepAlive to
be on.
Hope this helps,
Best regards,
Laurent Dami
I've had weird errors that look like a DNS error from IE before.
Something along the lines of 'the site is unavailable'. This was
infact the error of IE not liking the way i was serving the static PDF
errors.
Can you tell us the exeact wording of the error dialog IE shows?
-ash
_______________________________________________
List: [email protected]
Listinfo: http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst
Searchable archive: http://www.mail-archive.com/[email protected]/
Dev site: http://dev.catalyst.perl.org/
