[Catalyst] Re: Delays in Explorer connections via Catalyst

[will trillich]

Ooh, ngrep is wicked awesome! Thanks for that tidbit--

Now, back to the salt mines: Adding --fork or --keepalive to the
command-line for the test server didn't make any improvement. Same problem
seems to happen (off and on) when running under FastCGI in Apache as well.

Using *ngrep* on the server, we can see the requests arrive at the server
immediately, as suspected:

T 2012/03/18 *13:56:27.290616* 192.168.0.119:58316 -> 192.168.0.10:3000 [AP]
  GET /incident/2499/tab/info?_=3D1332096987014 HTTP/1.1..x-requested-with:
XMLHttpRequest..Accept-Language: en-us..Referer:
http://localhost:3000/incident/2499/..Accept: */*..UA-CPU:
x86..Accept-Encoding:
   gzip, deflate..User-Agent: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT
5.1)..Host: localhost:3000..Connection: Keep-Alive..Cookie:
incident_web_session=3D48e12cad549328590ac73b82396f01376129de24...
  .
#####
T 2012/03/18 *13:57:45.686752* 192.168.0.10:3000 -> 192.168.0.119:58316 [AP]
  HTTP/1.0 200 OK..Cache-Control: no-cache..Connection: close..Date: Sun,
18 Mar 2012 18:57:45 GMT..Content-Length: 9616..Content-Type: text/html;
charset=3Dutf-8..Set-Cookie: incident_web_session=3D48e12cad54932
  8590ac73b82396f01376129de24; path=3D/; expires=3DSun, 18-Mar-2012 20:57:45
GMT; HttpOnly..Status: 200..X-Catalyst: 5.80032....
#

Here you can see there's a MINUTE AND 18 SECONDS between request @13:56:27
and response @13:57:45! Argh! Once it does make its way back to the browser
everything is golden. But making users wait for 75+ seconds just isn't
feasible.

We're testing IE7 since that's what our client's clients seem to use :(
unfortunately.



So there's frequent (but not 100% consistent) delay on the server-side when
an Explorer request comes in to Catalyst, which doesn't affect FireFox or
Chrome. Brainstorming welcome! Ideas?



On Sun, Mar 18, 2012 at 1:01 PM, Adam Sj=F8gren <asjo at koldfront.dk> wrote:

> On Sun, 18 Mar 2012 11:37:44 -0500, will wrote:
>
> > When we run the server via "*script/*server.pl -d*" we sometimes don't
> even
>
> What happens if you add '--fork' to the command? Or perhaps
> '--keepalive' (I seem to remember IE liking that, at least at one point
> in time)?
>
> Could it be a Content-Length issue? (The browser waiting for the number
> of bytes in the header, the number being off (too high)?)
>
> I can, by the way, recommend ngrep=B9 for debugging stuff like this, it's
> quite nice.
>
>
>  Best regards,
>
>    Adam
>
>
> =B9 http://packages.debian.org/sid/ngrep
>
> --
>  "500 mil i sn=F6n                                              Adam Sj=
=F8gren
>  Ett UFO =F6ver sj=F6n                                      asjo at koldfron=
t.dk
>  och i en blixt dansar
>  Lady Day vals med Astaire"
>
> _______________________________________________
> List: Catalyst at lists.scsys.co.uk
> Listinfo: http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst
> Searchable archive:
> http://www.mail-archive.com/catalyst@lists.scsys.co.uk/
> Dev site: http://dev.catalyst.perl.org/
>



-- =

"We act as though comfort and luxury were the chief requirements of life,
when all that we need to make us happy is something to be enthusiastic
about." -- Albert Einstein
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.scsys.co.uk/pipermail/catalyst/attachments/20120318/1259b=
be3/attachment.htm