You are not logged in. Please login or register.
If there really is an issue in the network transition between EU countries, AC could ask thier ISP to prioritize some BGP routes to the US differently, but it would have to be a systemic issue with that NAP to make any noticable difference.
Usual it goes like this.
Lag...Freeze...DC
or Lag...DC
strange things have happened to: someone asked me once where I was going, I had dced, told them I had no idea, yet my bot was just walking on its own and then stopped, almost like the lag or some packet had caught up to server telling it I wanted to move.
My average ping is aroudn 270-350, which for me is playable. Lag spikes that kill hit around 700+, I know a dc is inbound from there, or it will creep up to about 500 or so and randomly freeze and go into a dc.
Yeah, I've ran PingPlotter through various disconnects of mine, and it always comes back to the same issue. The thing is I RARELY get any lag. I can be running through the island at a cool 150-250ms, then you get the tell-tale sign of a disconnect -- all of your actively running modules "freeze" and everything else stops moving. Then a disconnect is about 5-10 seconds away. I never get any lag beforehand. It's always just random.
After having run PingPlotter through several of these disconnects, as I stated, comes to the same issue. When I disconnect, I get 100% packet loss through: po5.ar2.FRA4.gblx.net (IP: 67.16.133.234)
I had initially posted that I thought this was a node in France, but a corp mate advised me it would be Frankfurt. Either way, that's the node (I guess that's the term) that is giving me trouble and the disconnects every time. I ran PingPlotter through 5-10 disconnects while in the background, and that was the one that showed the 100% packet loss through those times, every time.
If you all need more information from it, let me know, but that appears to be my issue. Arga/TBA -- have you guys run PingPlotter while you were lagging or disconnecting to know which is the problem for you?
I have not, but its probably a good idea, if I can verify that this node is my issue too, than thats something an ISP can work with to route our IP blocks through a different AS... Or the EU equivalent, they should be using BGP but I admit I don't do any overseas router work 
Well, it seems to get worse by the day for me. Every day, it's at least 3 times an hour. Just now, I woke up and logged on -- I have disconnected twice in 10 minutes from the game and was running PingPlotter during both times. Still the same culprit as above.
I'm willing to try almost anything. Is the only solution to try to get my IP to route through a different international node? Or is there anything else I can do? I really enjoy playing PO and have invested a lot of time into it since December, so I'd hate to just up and leave, but you can't play the game disconnecting every 5-20 minutes...hell, when it was bad, I used to just mine and deal with the disconnects but even that is difficult to do now.
Any and all suggestions welcome. And before anyone asks, yeah, I've tried resetting my router, unplugging my modem and let it sit overnight disconnected, etc. Once again, nothing else using my internet is ever affected, so I really don't think it's any problem local, but I'm willing to try all suggestions that anyone can come up with.
After having run PingPlotter through several of these disconnects, as I stated, comes to the same issue. When I disconnect, I get 100% packet loss through: po5.ar2.FRA4.gblx.net (IP: 67.16.133.234)
how bout that, same for me as well. Where you from Lemon?
I'm from Texas. I saw a few posts back you were from Idaho but, yeah, if Arga (And others who may not be reading this thread) have the issue with this same router, is it possible for the PO devs to do something on their side (With their server, contacting their IP, or something else)?
I hate to put this off onto them as it's not their problem, but I really doubt my IP would change their routing around it for just one customer, heh.
According to http://www.bgp4.as/internet-exchanges there is only one peering point (exchange) out of Hungary.
FRA4.gblx.net is part of AS 3549 (owned by global crossing)
That's good news and bad. The bad news is if your ISP is peering with global crossing then you're going to go over that node (it looks like its fraNAP in germany with the issue). The good news is it isn't the only node. The bad news is that most ISP's in the US peer with gbx unless they have thier own overseas fiber.
From work here, we use Cogent and they peer directly into Budahpest. But I have cable at home, which peers to gbx.
Moving the server to someplace with more than one exchange would be the best solution.
Edit: just wanted to add that cogent goes through fraNAP too, so the issue is probably with gbx's router and not the fraNAP/BIX link.
So, basically, it's not looking good? Hmm, glad I didn't get that 4th account. I'll wait and see a bit longer to see what happens, I suppose. Thanks for the information.
Sent a mail to our provider, let's see how far that goes.
It couldn't hurt ... because its like totally digital mail, unlike real mail which can cause paper cuts, infections, and eventual amputation!
I have been having lags and disconnects since T.I. was released. Defragged, then erased and reloaded both clients.
This is what I saw May 9th between 6:30p and 7p EST.
http://img607.imageshack.us/img607/4396 … 007312.png
As I read it, It shows that the connection is good until it gets to Budapest, then shows the Perpetuum server was unreachable from there. Anyone please correct me if you see something different, but it appears the problem I am having is on the server side.
brand new puter i7 just started playing last week, very high d/c occurances. I'm located in Abilene, TX my ip is Suddenlink.
This is where the problem starts. any more suggestions?
9 207 ms 201 ms 171 ms po5.ar2.FRA4.gblx.net [67.16.133.234]
10 189 ms 187 ms 187 ms 204.245.36.142
11 187 ms 194 ms 187 ms 81.183.0.78
12 187 ms 185 ms 185 ms 81.183.0.35
13 189 ms 187 ms 187 ms 81.183.2.225
14 192 ms 190 ms 194 ms 84.2.225.113
15 189 ms 189 ms 189 ms gameserverb.perpetuum-online.com [195.228.152.152]
Trace complete.
C:\Users\Stormtrain>
C:\Users\Stormtrain>
This is where the problem starts. any more suggestions?
9 207 ms 201 ms 171 ms po5.ar2.FRA4.gblx.net [67.16.133.234]
10 189 ms 187 ms 187 ms 204.245.36.142
11 187 ms 194 ms 187 ms 81.183.0.78
12 187 ms 185 ms 185 ms 81.183.0.35
13 189 ms 187 ms 187 ms 81.183.2.225
14 192 ms 190 ms 194 ms 84.2.225.113
15 189 ms 189 ms 189 ms gameserverb.perpetuum-online.com [195.228.152.152]Trace complete.
Without seeing the hops before we can't tell if it's a distance or router issue. If you're in the US for example and hop 8 was to somewhere in the US you're crossing an ocean in one hop that's going to add some latency. If however you're in France (Location of this router) and seeing this level of MS then something is wrong.
What you're looking at is pretty much the average of the US Latency.
Alex, the router is in FRAnkfurt, and its the same place where many others are seeing an issue. This is the 'landing' point for global crossing before the signal gets routed into Hungary's single Network Access Point, a public peering point in germany.
Any word back from the ISP?
They basically said they don't have any influence on international networks and can't redirect traffic...
Ok, not an unexpected response, thanks for trying.
http://www.cogentco.com/files/images/ne … _large.jpg
As you can see on the map, US-Frankfurt is the shortest hop, then a direct fiber link down through munich/vienna/bratislava to budapest.
Technically, they could probably prioritize traffic to go out through Timisoara, but that's really a dead end and the traffic would just end up going north anyway.
Note however, all the pretty blue lines in/out of London both to EU and to US
Is moving the server to a location which has more flexibility (US, UK, etc) out of the question?
it certainly isn't a simple as me typing "move server".
I suspect the logisitcal, legal, and financial aspects of such a move will need a serious level of review before they can even consider it.
I pinged and did a tracert on the above ip's noted in my post. all functioned correctly (dallas, washington, frankfurt, budapest) except for one. This one 204.245.36.142 .
The ping times out, the tracert times out. It shows to be in Phoenix, AZ. USA which is where GBLX is located. You cannot retrieve a hostname or domain for this ip. don't know if that means anything but there is something definitely wrong with this ip.
100% loss everytime.
I don't know too much about all this but this ip does not act like the others.
I get this as well but only if I attempt to play during the US daytime. In the evenings it works like a champ. Without looking into it it makes me wonder if congestion on the gbx pipe during prime business hours is what the problem may be and why not everyone is complaining in the US?
As an aside, long term I think we need to figure out a better datacenter solution that has some stronger peering options (read: colo.) I understand that with a small team this creates some logistical issues but a plane flight once a month for maintenance and remote support for rebooting incase of a failure shouldn't be an unrealistic way to accomplish this. The discussion should be around how do we get there not why we can't.
The short term fix is for all us US whiners to get a real job and stop trying to play from 8-5
RG_ not all equipment will answer ICMP:Echo requests, even if they are alive and online. ICMP:Traceroute is a different code. http://en.wikipedia.org/wiki/Internet_ … e_Protocol
If you can ping the IP address AFTER that one, it means the packets are going through the unpingable device just fine.
Glimpse, This happens to me on the west coast 9pm-12 PST, but I had a dream I was fired last night, so maybe I'll get to play during the day too
Powered by PunBB, supported by Informer Technologies, Inc.
Currently installed 3 official extensions. Copyright © 2003–2009 PunBB.
Generated in 0.052 seconds (83% PHP - 17% DB) with 21 queries
Copyright © Avatar Creations
