EAB Final Update

[prowler]

Originally Posted by Ed Cruse

This whole problem started when I switched to a new computer with W7 and IE10, my old computer used XP with IE8, so I assumed the problem was here. IE10 in IE8 mode does the same thing so I'm going to fire up my old computer and see if it works or not. I'll get back with the results.

I am using IE8 in WinXP at this moment, and I'm having no problems whatsoever with loading pages.

[RCK]

Originally Posted by Ed Cruse

I wonder how good the IE8 mode on IE10 actually is.

One kind of emulation, not the real thing. (c)Akira
Can you turn off anti-virus / anti-malware, or any other kind of network protection (peerblock, etc.) ?

[Ed Cruse]

Originally Posted by RCK

One kind of emulation, not the real thing. (c)Akira
Can you turn off anti-virus / anti-malware, or any other kind of network protection (peerblock, etc.) ?

All that has already been turned off. The only thing that I have running is W7's FireWall.

From this end the data is simply not getting here without the occasional delay, the traffic capture shows a long delay on the last line and watching the modem lights confirms that. What's strange is the traffic capture shows as much as a 60s delay and the web page shows a 0.3s delay. It kind of makes me wonder if there's a thread collision in the code on the server that measures the time it takes to send a chunk of data. This could be causing the delay and causing the timer to start timing after the delay occurs.

I don't think the delay is between us because a whole bunch of data comes in immediately after clicking with little delay until the last line. I assume each time the server is going to send data it tells IE in which case IE starts it's timer and then stops the timer when the data has been sent. This delay doesn't happen every time, it seems to be a lot worse in the morning here and best in the afternoon here. You're about 8 hours ahead so the worst times are in the late afternoon there and the best times are late at night there. I assume server activity is highest in the late afternoon there.

Normally I would blame my end for strange problems like this, but I've had no problems with any other web page except ones that have "abime.net" in their addresses. I assume those web pages are on your server.

[lilalurl]

Originally Posted by Ed Cruse

I don't think the delay is between us because a whole bunch of data comes in immediately after clicking with little delay until the last line.

Back when I was active in the UT2004 community, server admins asked players who had packet loss issues to run WinMTR for a while (at least 30 minutes) and post the result:
http://winmtr.net/how-to/

I remember that it helped to discover an issue at the node between a German ISP and the UK network.


Not sure if it would serve purpose here, but your case reminded me these issues.

RCK will probably be able to say if it would be useful or not.

[RCK]

it's a good idea lilalurl, if you could run winmtr with one abime.net filter during 30 min of navigation, we would be able to know if it's a network problem.

[Ed Cruse]

Originally Posted by RCK

it's a good idea lilalurl, if you could run winmtr with one abime.net filter during 30 min of navigation, we would be able to know if it's a network problem.

I'll down load Winmtr and read the directions. At this point I'm not sure what you mean by one abime.net filter, hopefully I'll figure it out.

[lilalurl]

Just follow the instructions of Step 3 and use abime.net instead of google.com (example given in the how to).

By filter, I think RCK meant the dropdown box in which to type the address.


I just learnt something interesting by doing the test myself.
It seems that to reach the abime.net server (North of France), my ISP network send the packets to Lyon (Lyons in English) which is North of me (and South of the abime.net server). But then it goes back to Marseilles (very South of France and South of me)



And there is a rivalry between the two cities about which one is the second major city in France. Seems like the network structure of my ISP has decided .

[Ed Cruse]

Originally Posted by lilalurl

Just follow the instructions of Step 3 and use abime.net instead of google.com (example given in the how to).

By filter, I think RCK meant the dropdown box in which to type the address.


I just learnt something interesting by doing the test myself.
It seems that to reach the abime.net server (North of France), my ISP network send the packets to Lyon (Lyons in English) which is North of me (and South of the abime.net server). But then it goes back to Marseilles (very South of France and South of me)



And there is a rivalry between the two cities about which one is the second major city in France. Seems like the network structure of my ISP has decided .

Thanks for the info. Interesting how packets don't take the short route and go the long way.

[Ed Cruse]

Originally Posted by lilalurl

Just follow the instructions of Step 3 and use abime.net instead of google.com (example given in the how to).

By filter, I think RCK meant the dropdown box in which to type the address.


I just learnt something interesting by doing the test myself.
It seems that to reach the abime.net server (North of France), my ISP network send the packets to Lyon (Lyons in English) which is North of me (and South of the abime.net server). But then it goes back to Marseilles (very South of France and South of me)



And there is a rivalry between the two cities about which one is the second major city in France. Seems like the network structure of my ISP has decided .

I've attach the results of the ping test. The first one was during a delay and the second one was after the delay. The first line is my router which the number of pings keeps accumulating, which I don't understand because all the other lines seem to show only the results for each test. The second line I think is my DSL modem which shows a high loss. I checked the modem and according to it for over an hour the DSL communications was perfect. No bad packets or dropped packets.

The only thing I noticed different between the two tables is during the delay the program window didn't blink as it updated, but the numbers changed. After the delay the program window blinked during each updates.

How does the program work? Does it send one ping at a time and each node response to it and then sends the ping on to the next one? Or does it send one ping at a time to each node that it finds? Judging by the numbers I'd say one ping hits all the nodes.

[lilalurl]

I don't have enough technical knowledge to give you an answer, but here is what MTR is:
http://en.wikipedia.org/wiki/MTR_%28software%29

RCK will probably be able to tell you more.

I only have experience from reading what server admins said but having some loss does not mean an issue per se because some routers are configured to ignore some requests (always or at a certain interval).
The interesting thing for admins is when there is a huge variation from one router to another (especially when it belongs to to the same provider).

[RCK]

There is too much packet lost on the two analyses to conclude anything
Hum, I don't have any more idea, can you make more IE10 network capture and send me them all when delay is occurring ? (And please confirm the server generation time at the bottom of the page is OK)

[Ed Cruse]

Originally Posted by RCK

There is too much packet lost on the two analyses to conclude anything
Hum, I don't have any more idea, can you make more IE10 network capture and send me them all when delay is occurring ? (And please confirm the server generation time at the bottom of the page is OK)

Here's another capture. When I clicked a link it showed one line and didn't do anything. I clicked it again and nothing showed until the end of the delay in which the stuff in the attached file showed all at once. I don't know if what I see is live or shows only after the last line comes in. I don't know how the capture works. The very first line I mentioned at the top of this paragraph disappeared. I thought the first line would be at the top with everything else appended to it, but wasn't.

Edit 1: At the bottom of the page it showed approximately 0.2s.

[Ed Cruse]

I captured several more. The second upload is interesting, I put some commentary inside it, it was from clicking "support.winuae". Every captured page in both files shows 0.3s or less.

I've noticed a lot of stuff that looks like tracking information, is that through your system or my system?

[Bamiga2002]

I just screengrabbed this a.ka "Spot the difference"
I'm using SRWare Iron (Chromium-based) and I was simultaneously online and offline(?)

[Lonewolf10]

Not sure if this is because the picture has no uploader named underneath, but I just joined the Pixel Pushers H.Q group. I have been able to view other pictures in the group, but this one...

http://eab.abime.net/group.php?do=pi...pictureid=1473

... comes up with an error. I was going to use the "administrator" link on the error page that appears, but that has the same link as above!

Is everything ok?? Is it possible to view that image?

[DH]

The images are uploaded via members albums, so my guess is, the person who uploaded that image has since deleted it from their album, which should have removed it from the group, but has for some reason left it :s

Or the member has been banned or pruned because of a low post count.

EDIT: I have removed the orphaned thumb image, obviously the database held onto the thumbnail, despite the image being deleted from the banned members profile. Anyway, it was not a problem with the vB update, it was more a problem with one of the members.

[Lonewolf10]

DH, thanks for looking into it. I'm glad it wasn't related to the update

[RCK]

Originally Posted by Lonewolf10

... comes up with an error. I was going to use the "administrator" link on the error page that appears, but that has the same link as above!

All those "Administrator" email link should be disabled, so it's normal it does nothing

[RCK]

Originally Posted by Ed Cruse

I captured several more. The second upload is interesting, I put some commentary inside it, it was from clicking "support.winuae". Every captured page in both files shows 0.3s or less.

I've noticed a lot of stuff that looks like tracking information, is that through your system or my system?

Interesting you have a lot of pending for awjs.js tracker (local abime.net)
I have temporary disabled awstats tracker that was (pending) into your IE10, please try to navigate again and tell me if it has fixed your IE10 freeze.

[Ed Cruse]

Originally Posted by RCK

Interesting you have a lot of pending for awjs.js tracker (local abime.net)
I have temporary disabled awstats tracker that was (pending) into your IE10, please try to navigate again and tell me if it has fixed your IE10 freeze.

Sorry, it didn't help. In fact right now it's really bad, it's 0820 here. Almost every time I click on anything I have to go to other web sites on another tab until IE stops waiting for abime.net.