Wireshark-users: Re: [Wireshark-users] TCP retransmit, HTTP and web page not loading
From: Steve Bertrand <[email protected]>
Date: Wed, 25 Jun 2008 10:38:09 -0400
Sake Blok wrote:
On Wed, Jun 25, 2008 at 09:19:46AM -0400, Steve Bertrand wrote:
There is a screen cap of the final communication between the server and client if anyone could possibly have a look if they can answer my questions.
http://ibctech.ca/screencap/fibre_communication.jpg
Please provide the (binary) capture file instead of a screendump. We love
using wireshark instead of an image-viewer ;-)
LOL, sorry. Like I said, new to list, and no charter or BCP ;)

That said, I certainly should have known better.

I hope this is what you are after. I'm not all that familiar with Wireshark, as I do all of my traces via tcpdump on FreeBSD. I just saved a new file from within Wireshark with the packets from the entire HTTP conversation in question:
http://ibctech.ca/http-conv.pcap

A) Will a TCP session finish successfully even if some frames could not be ultimately received?
Nope, TCP keeps state information on it's connection and tries to
recover from missed data and also when data has not been acknowledged.
Is this why sometimes a browser will just be apparently 'spinning it's 
wheels'?
Since I've had to delve into this the last couple days, I'm going to go 
off on a tangent to learn much more about this level of network 
communication.

Symptoms are the Title of the site appears in the browser, it spins it's wheels, then stops.
Taken that this particular object never makes it to the client, this is
to be expected as the object is a cascading stylesheet, which would be necessary to format the html-code on the screen. When it's missing, the
browser can not show the data it has already received.
This is exactly what I had thought.

Please upload the binary capture so we can have a closer look at
where the data might be missing.
Let me know if the file I've included the link for is what you need.

Next steps would be to start a capture on both ends of the fibre to see where there's a difference between the two captures.
This was our objective today.

Is the ethernet
data tunneled somehow by the fibre-provider?
Well, I don't believe so. I connect my switch to a switch at an 
intermediary, and the client end switch is connected to the same 
intermediary switch.
From what I can tell, there is no trickery involved. However, it is 
excruciatingly painful getting any decent answers from the intermediary, 
but we have no choice but to use their fibre (they are the PUC, and they 
own it). I'll only go to them when I can absolutely confirm that both my 
end, and my client's end are functioning correctly.
Thanks for the insightful feedback.

Steve

Cheers,
    Sake
_______________________________________________
Wireshark-users mailing list
[email protected]
https://wireshark.org/mailman/listinfo/wireshark-users