the diff.txt for IE was:
Comparing files begin5.txt and AFTER5.TXT
Let's summarize the non-zero differences for clarity:
(Search for prefix Net to find these summary additions.)
***** begin5.txt
Packets Received = 3625
Received Header Errors = 0
***** AFTER5.TXT
Packets Received = 3630
Received Header Errors = 0
*****
Net Packets Received = 5
***** begin5.txt
Received Packets Discarded = 0
Received Packets Delivered = 3625
Output Requests = 3660
Routing Discards = 0
***** AFTER5.TXT
Received Packets Discarded = 0
Received Packets Delivered = 3630
Output Requests = 3673
Routing Discards = 0
Net Received Packets Delivered = 5
Net Output Requests = 13
The delivered packets number seems consistent with the idea that
you are getting a single error message. I don't know what that high
number of output requests will mean. It could mean that that is the
number of HTTP request headers that are being sent (such as the
User-Agent: header I alluded to earlier). That is why it would be useful
I think to capture the actual packets and be able to format them for
further analysis. XP has a simple means of capturing them with its
netcap command. Unfortunately Microsoft only provides a formatter
for the .cap files that that creates with Windows server. That is what
Ethereal would do for you. (FYI Ethereal is third-party freeware.)
The number of packets and requests for the non-IE case seem
too large to be interesting. Did you try using telnet instead?
The results of that test should be interesting because if it works
it should show you a much smaller packet count than your alternate
browser test and also if it doesn't work it may show you exactly the
same result that IE is getting with much smaller packet and request
counts.
Coincidentally I was trying to help someone else with a related problem
yesterday and suggested the following:
<excerpt>
netstat -s >before.txt
telnet 207.46.244.188 80
<wait for the screen to clear and type GET /
then press Enter
netstat -s >after.txt
fc before.txt after.txt >diff.txt
notepad diff.txt
Here is a summary of that comparison (by subtracting the changed
statistics and reporting only non-zero results):
Packets Received = 36
Received Packets Delivered = 36
Active Opens = 1
Segments Received = 36
Segments Sent = 30
Note: for some reason my first attempt was dropped before I had
had a chance to type the GET command.
Make sure that you minimize the number of other Internet applications
to make your results comparable. Do more than one capture
if necessary to prove that your results are repeatable.
Note that this procedure bypasses the use of DNS.
Your ping test showed that your DNS seems to be working fine
but if the above test works try it again with the symbolic address.
In order to bypass the dnscache and ensure that you do a real lookup
you should first do:
ipconfig /flushdns
(After that I would also do:
ipconfig /displaydns | find /i "microsoft"
just to be sure that no unexpected entries are being added by your
HOSTS file.) Then if you do the same kind of comparison of before
and after statistics you will probably see some indication of the lookup
being done (e.g. with changes in Datagrams statistics).
If it turns out that these tests are successful we will have proved that
connectivity is not the problem. Therefore, it would have to be something
particular that IE and that site are both doing together.
</excerpt>
In your case Gemma you would substitute in the above telnet command
whatever address (symbolic or numeric) that you used for your tests.
(Or you could use the same address that the other user was trying.
It's just an alias of
www.microsoft.com.)
....
im not sure what ethereal is, and i dont know a lot about my
configuration, but if there is anything you need to know, then i will
find out for you. I used to have Norton installed, however i found
this slowed down my internet, so i turned it off. i was wondering if
this was the problem?
If you only "turned it off" or didn't successfully uninstall it completely
that could be very similar to the other user's problem that I mentioned.
Unfortunately I don't know anything about such products. I don't even
know if traces of them would have to appear in the Task Manager's
Processes tab but FWIW I guess that that is where I would start looking.
Thanks for the help. bye!
Good luck
Robert
---
Gemma Clark said:
....