Re: [analog-help] time out errors

2004-07-12 Thread analog-help
They don't show up at all, Kerstin

A 'timed-out' request (from the visitor's point-of-view) never actually
reached your server, or got a reply back to the visitor - that's why they
are seen by them as 'timed-out' (and not 'refused', or 'non-existent', or
some other error). So the server has no way of knowing they were ever made,
hence, cannot record anything to that effect.

The nearest you might see is an error 206, which means 'content incomplete'
or 'transfer incomplete' or similar. I get this occasionally on a big WMV
video file tutorial on my site, which sometimes people watch half-way
through, and then cancel the viewing at their side. In those circumstances,
because the cancellation emanates from their side, and is responded to (and
cancelled) by my server, it CAN record a 206 error.

But timeouts are different. Think of them instead, as rather like How many
visitors (that you didn't expect to be popping round) weren't actually able
to find your house or street? There is, after all, no way you could know,
is there?

Hope this helps,
Regards
Neil

- Original Message - 
From: Kerstin [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Monday, July 12, 2004 7:30 PM
Subject: [analog-help] time out errors


 Hi:

 Where would the timed out requests show up?

 Would it be under the failure report?

 Thanks, Kerstin



 +
 |  TO UNSUBSCRIBE from this list:
 |http://lists.isite.net/listgate/analog-help/unsubscribe.html
 |
 |  Digest version: http://lists.isite.net/listgate/analog-help-digest/
 |  Usenet version: news://news.gmane.org/gmane.comp.web.analog.general
 |  List archives:  http://www.analog.cx/docs/mailing.html#listarchives
 +



+
|  TO UNSUBSCRIBE from this list:
|http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
|  Digest version: http://lists.isite.net/listgate/analog-help-digest/
|  Usenet version: news://news.gmane.org/gmane.comp.web.analog.general
|  List archives:  http://www.analog.cx/docs/mailing.html#listarchives
+


Re: [analog-help] time out errors

2004-07-12 Thread analog-help
On Mon, 12 Jul 2004, Neil D. Jackson wrote:

 They don't show up at all, Kerstin

 A 'timed-out' request (from the visitor's point-of-view) never actually
 reached your server, or got a reply back to the visitor - that's why they
 are seen by them as 'timed-out' (and not 'refused', or 'non-existent', or
 some other error). So the server has no way of knowing they were ever made,
 hence, cannot record anything to that effect.


That's true, but there is an HTTP code 408 Request Timeout, if the server
gives up waiting for the client. I can't work out when this comes about
though.

 The nearest you might see is an error 206, which means 'content incomplete'
 or 'transfer incomplete' or similar. I get this occasionally on a big WMV
 video file tutorial on my site, which sometimes people watch half-way
 through, and then cancel the viewing at their side. In those circumstances,
 because the cancellation emanates from their side, and is responded to (and
 cancelled) by my server, it CAN record a 206 error.


No, 206 doesn't mean that. It means that the client requested only part of
the file, and the server successfully delivered the range that was asked
for.

-- 
Stephen Turner, Cambridge, UKhttp://homepage.ntlworld.com/adelie/stephen/
  Low Priced Cambridge Clare College. Big selection at eBay UK!
  (Ad after Google search for Clare College Cambridge)
+
|  TO UNSUBSCRIBE from this list:
|http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
|  Digest version: http://lists.isite.net/listgate/analog-help-digest/
|  Usenet version: news://news.gmane.org/gmane.comp.web.analog.general
|  List archives:  http://www.analog.cx/docs/mailing.html#listarchives
+


Re: [analog-help] time out errors

2004-07-12 Thread analog-help

- Original Message - 
From: Stephen Turner [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Monday, July 12, 2004 11:21 PM
Subject: Re: [analog-help] time out errors


 That's true, but there is an HTTP code 408 Request Timeout, if the
server
 gives up waiting for the client. I can't work out when this comes about
 though.

Hmm, not sure - can't recall ever seeing one myself, as it goes. Maybe this
could be caused by a 'Connection: Keep-alive' HTTP from the client, which
then isn't kept open for successive requests? Or more likely, when the
server is 'chunking' data to the client, and it dies halfway through? Not
sure.


 No, 206 doesn't mean that. It means that the client requested only part
of
 the file, and the server successfully delivered the range that was asked
 for.

Ah, right - LOL - I'll shut up then :) Thanks for the explanation.

Just out of interest, though, how could a client ask for 'part' of a WMV
file? All it can ask for, is the complete URL, surely? Or is there some
kind of 'resume' operation available in HTTP as there is in FTP? That's
kinda why I read between the lines, and put it down to the client closing
the data-stream before it was finished, hence the server recorded it had
sent what it had been asked for, up to that point.

Actually, I'm more confused now than I was to begin with - LOL.

Cheers
Neil


+
|  TO UNSUBSCRIBE from this list:
|http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
|  Digest version: http://lists.isite.net/listgate/analog-help-digest/
|  Usenet version: news://news.gmane.org/gmane.comp.web.analog.general
|  List archives:  http://www.analog.cx/docs/mailing.html#listarchives
+


Re: [analog-help] time out errors

2004-07-12 Thread analog-help
On Mon, 12 Jul 2004, Neil D. Jackson wrote:

 Just out of interest, though, how could a client ask for 'part' of a WMV
 file? All it can ask for, is the complete URL, surely? Or is there some
 kind of 'resume' operation available in HTTP as there is in FTP? That's
 kinda why I read between the lines, and put it down to the client closing
 the data-stream before it was finished, hence the server recorded it had
 sent what it had been asked for, up to that point.


It can certainly ask for part of it if it wants -- there is a syntax for
that in HTTP. This is common for PDF requests -- many clients ask for them
a page at a time, because you often don't end up reading the whole document.

-- 
Stephen Turner, Cambridge, UKhttp://homepage.ntlworld.com/adelie/stephen/
  Low Priced Cambridge Clare College. Big selection at eBay UK!
  (Ad after Google search for Clare College Cambridge)
+
|  TO UNSUBSCRIBE from this list:
|http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
|  Digest version: http://lists.isite.net/listgate/analog-help-digest/
|  Usenet version: news://news.gmane.org/gmane.comp.web.analog.general
|  List archives:  http://www.analog.cx/docs/mailing.html#listarchives
+


Re: [analog-help] time out errors

2004-07-12 Thread analog-help

- Original Message - 
From: Stephen Turner [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Monday, July 12, 2004 11:56 PM
Subject: Re: [analog-help] time out errors



 It can certainly ask for part of it if it wants -- there is a syntax for
 that in HTTP. This is common for PDF requests -- many clients ask for
them
 a page at a time, because you often don't end up reading the whole
document.


Ah, yes - makes perfect sense now - thanks, Stephen. I did a spot of RTFM
too - I see it now: Content-Range header of the HTTP request.

While I was at it, I looked up 408 too - the definition given in my book is
the server timed out waiting for the full client request - so I'm
guessing that this would only occur in situations where the client passed
only some of its HTTP request headers, and didn't manage to reach the stage
of the double CRLF before the server decided to give up. A fairly rare
situation, I'd guess, but possible, so it needed a status code.

Cheers
Neil


+
|  TO UNSUBSCRIBE from this list:
|http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
|  Digest version: http://lists.isite.net/listgate/analog-help-digest/
|  Usenet version: news://news.gmane.org/gmane.comp.web.analog.general
|  List archives:  http://www.analog.cx/docs/mailing.html#listarchives
+