5 replies to this topic

[Grifthin]

I've been monitoring MWO for a while now and I think I've figured out why some people DC or crash intermittently. Occasionally MWO will use ALL available bandwidth to download/upload (I don't know too what) which causes your ping in game to soar before crashing.

It uses the following ports:
70.42.29.75 port 21908
70.42.29.75 port 21325
23.198.49.103 port 80
70.42.29.74 port 45461
Shaping the connection with something like Netmonitor causes the game to be unplayable. All you can do is wait it out and a few minutes later you can once more play. But even if you restart your client it has no effect. You can do nothing but wait till it's done. It's almost like their pulling some kind of data from your client. No other processes were using data and I tried this on multiple PC's from different sites with various connections to test. isn't Specific to my computer.

EDIT: Screenshot. Also seems only to happen while dropping 4 mans. My maximum download speed is 30kbps.

Edited by Grifthin, 12 September 2013 - 10:19 AM.

[Shamous13]

Interesting find, I'll see what my connection is like the next few days.

[TheFlyingScotsman]

I can collaborate this. My crashes always come in bursts, and usually start with a mid-game crash followed by crashing during any attempt to play, at the start of the match.

I don't know if this is from PGI data mining or ghost viewing, but it is unavoidable once it starts and is only playable once it ends, even through computer restarts.

PGI, this thread needs your attention, please look into the matter, and more importantly, be honest with what you find.

Edited by TheFlyingScotsman, 12 September 2013 - 11:27 AM.

[Deathlike]

MWO tends to "stay connected" despite your client crashing. I don't quite understand why, but it is somewhat linked to how long your mech is "held up". I get the feeling the server doesn't notice or detect that, despite the port staying open.... and you continue to be flooded with data that you can't use because MWO has crashed.

[DragonsFire]

Definitely some interesting finds, and while I agree that the thread needs attention and further corroboration by others, have any of you opened a support ticket with the findings as well?

That is a good way to get the ball rolling and will give them further info on where to start and what factors to look for as unfortunately thread troubleshooting sessions can end up being a bit nebulous and cumbersome. That's not to say that they aren't valuable by any means either, just that getting a ticket started helps the whole process along.

[Grifthin]

I opened up a support ticket. Hopefully this helps them find this bug.