Curse

Addone_Abaddon · Sep 02, 2006, 08:10 PM // 20:10

I got error 007 in the middle of HA, lucky it was HA with my guild.

Eviance · Sep 02, 2006, 10:36 PM // 22:36

Quote:

Originally Posted by johnmedgla

Thankyou God for Servants of Fortuna, and thankyou expecially Heist23 for pointing it out - the Sysinternals Defrag tool mentioned above worked like a charm. 200 Fragments per file down to 1, and the Lag has vanished! I've played for three hours now without any lag or disconnects, which is a minor miracle lately, given that I was disconnecting every half hour or so before. THANKYOU!!!!

If you missed Heist23's post, I'll copy the link the he provided (hope you don't mind)
http://forum.sof-guild.com/showthread.php?t=5100

I wish it had worked for me - I had one file to defrag and that was it -_- Meaning that again mine must be elsewhere in the mess.... I still need to try tracert - I guess I will log on now to try that -_-

Errm I did the "C:\Program Files\Guild Wars\Gw.exe" -diag ... I would so post what it said if I could read the damned thing >_> It looks like a ton of pretty numbers to me >_< (I'm sure given some time and rest I could get it, but I had a long day at the dress shop and numbers are the last thing I want to look at right now).

Gaile Gray · Sep 03, 2006, 01:30 AM // 01:30

Quote:

Originally Posted by seut

Problem seems to be the routing from Limelight Networks. Sometimes when lagging out you get an IP in the error msg. I did a traceroute and got the following:

$ tracert 216.107.240.197
routing to 216-107-240-197.plaync.com [216.107.240.197]
1 8 ms 9 ms 8 ms (censored my local german provider)
2 8 ms 8 ms 8 ms
3 18 ms 17 ms 17 ms
4 18 ms 17 ms 17 ms (major european internet hub)
5 33 ms 39 ms 33 ms ve1317.ge1-9.fr1.ams.llnw.net [80.81.192.221]
6 29 ms 27 ms 28 ms tge1-16-1.ar1.ams.llnw.net [69.28.172.86]
7 248 ms 97 ms 97 ms so1-5-1.ar1.lga.llnw.net [69.28.172.161]
8 102 ms 128 ms 96 ms tge1-4.fr3.lga.llnw.net [69.28.172.2]
9 129 ms 129 ms 128 ms tge2-1.fr3.ord.llnw.net [69.28.171.193]
10 169 ms 168 ms 169 ms tge1-2.fr3.sjc.llnw.net [69.28.171.66]
11 177 ms 178 ms 194 ms tge2-3.fr3.lax.llnw.net [69.28.171.117]
12 timeouts from here on

My err7 disconnects coincide with the disconnects of guildmembers in the US doing other stuff in GW and random people in my pugs.

Ok, does that or does that not show it's not Guild Wars, but instead is the provider or the backbone over which the data travels? Or the provider? I just glanced at the trace, but honestly, so many many times, we look at the records and the problem is external to Guild Wars. Nobody wants to believe that, they think we're trying to cover up game errors. But in fact more often than not, it's getting to us, and not within our system, that is the major problem. We cannot reach out and nab your connection -- it has to go through all the jumps. And when you see timeouts in the jumps, then you really cannot honestly blame Guild Wars.

I am sorry if I sound like I'm making excuses, but that fact is, I'm really sad that people keep insisting we "fix the problem!" when the problem sometimes lies outside our network.

Quote:

Originally Posted by Curse You

I forgot to point out something before. I WAS having lag in Guild Wars on my older computer, but I have somehow managed to reduce it down to almost nothing. The only things I have done are:

1. Defraged
2. Re-installed Video and Sound Drivers, and updated Direct X
3. Moved a bunch of games to my secondary hard drive, to giving me 16 GB of free space
4. Put in a new modem
5. Cleaned my PC (the physical unit)

Somehow, doing those things has helped to reduce my lag in game. The most likely thing that did something was making 16 GB of free space and putting in a new modem.

Yes, all non-Guild Wars fixes.

Quote:

Originally Posted by samifly

For all of you who arent suffering from lag, great, congratulations, you are one of the lucky few. Half of the community is suffering from this lag. This is not, a computer issue, if 4 people disconnect at the same time.

Are they nearby? Do they use one of the few backbones into the game servers? Could be us, yes, and could be external to us. Fill out a support ticket. Send in the requested data that you obtain from one of two programs. They will help you, if they can, if it's Guild Wars that is the problem. Guild Wars uses more of your system requirements than most other applications, so you cannot say "Everything else is fine, it's the game!" In fact, it may still be the computer, a computer that is struggling to run a game that wants to make the most of the resources available to it.

Quote:

Originally Posted by Inde

But see Burn, this proves one of my points exactly. The network engineers, the server administrators (whatever title they give themselves) have probably done the same as you and automatically written off this problem as belonging to the user so what's the point of checking into it. I work with a lot of these people and I can tell you firsthand that this is the first response of a server admin. Only until they can see it affecting many people will they even take the time to research into it. The fact that, again, SO MANY have been reporting this problem as of late means that something is indeed going on that is out of the user's control and does indeed point to ArenaNet.

Actually, we ask that people run a utility or two and send us the resultant reports. We don't mind if it's one person or one hundred, we really do want to help, and those reports provide the means to do so, as they help us pinpoint the source of the problem. If we need to tweak programming we will. If we need to ask others to tweak theirs, we will.

Quote:

Originally Posted by Carinae Dragonblood

Even if it isn't their fault, they need to publically acknowledge the problem and help locate it! I'm sure Anet can put more pressure on a backbone provider than you or I can.

Don't we wish! Are can put some pressure on our ISPs and so forth, but I don't think many of the major backbone providers are too sympathetic to our concerns. I could be wrong -- maybe they're as responsive as Mary Poppins. But my overall feeling is that they don't exactly hop to help us. We really do want to work on this to the best of our ability and will do what we can when it's internal to our data centers, and will work with our providers to affect improvements, too. We just need that information, that proof of where the problem is so we can fix it or say "See, here's the bug in your system" and get it fixed!

Quote:

Originally Posted by EternalTempest

I did not get a chance to send the diag report to anet last night, but will do so today.

Thank you very much!! If it's us, we'll work on it. If it's someone we're paying, you can be sure we'll lower the boom!

Carinae · Sep 03, 2006, 02:01 AM // 02:01

I'm just saying the symptoms are giving the distinct impression of being upstream from the end-users.

A wide geography is affected, ruling out (most likely) individual PC issues. Users located in other areas, regionally located, are experiencing no issues.

There might be a clue to this in that.

The problem began at the start of the previous weekend event, before the GWFC.

There might be a clue to this in that.

Multiple people are often affected at the same time. Partied or Unpartied, Different Chapters, Different Guilds.

There might be a clue to this in that.

Things are pointing upstream at either a backbone issue, or an Anet datacenter issue. Hopefully, it'll be found soon, where ever it is.

MerLock · Sep 03, 2006, 02:25 AM // 02:25

Haven't had many problems with lag but have a lot of Error 7 disconnects in the past few days. I've had about 6 in a period of 2 days and before that, I've only had 2 error 7 since I bought the game. And I bought Guild Wars when prophecies first came out!

Curse You · Sep 03, 2006, 02:29 AM // 02:29

This was pointed out before, but, the problem could very easily be on your computers, all of your computers.

This thread has some reasons why:
http://www.guildwarsguru.com/forum/s...php?t=10037429

I suggest EVERYONE (even those with no problems) Defrag their Gw.dat file, as shown in the thread.

Gorebrex · Sep 03, 2006, 02:55 AM // 02:55

Quote:

Originally Posted by Curse You

This was pointed out before, but, the problem could very easily be on your computers, all of your computers.

This thread has some reasons why:
http://www.guildwarsguru.com/forum/s...php?t=10037429

I suggest EVERYONE (even those with no problems) Defrag their Gw.dat file, as shown in the thread.

Thanks to Gaile for telling us a little more of whats going on on ANets end of it!

Heres my tracert to the same IP Seut used

Its pretty obvious its NOT my PC, but a ways down the line where the trouble starts. Sure defragging might help, but wont get me a shorter hop.
Its a little better now

samifly · Sep 03, 2006, 03:13 AM // 03:13

They are from multiple parts of the world Gaile. I've filed a support ticket, which seemed to go bust, and the whole "people on opposite sides of the continent disconnecting at the same time" was ignored by PlayNC. And I am fairly fluent in computers, so I am 99.9876% sure its not my system.

Quote:

Originally Posted by Curse You

This was pointed out before, but, the problem could very easily be on your computers, all of your computers.

This thread has some reasons why:
http://www.guildwarsguru.com/forum/s...php?t=10037429

I suggest EVERYONE (even those with no problems) Defrag their Gw.dat file, as shown in the thread.

I'm sorry, but the likelyhood of everyones gw.dat file being corrupted at the same time, is much less than there being a network problem on GW's end. And yes, I tried that. Still lag, still error 7ns, and it was only 80 files fragmented.

gr3g · Sep 03, 2006, 03:56 AM // 03:56

I love seeing people who have no clue about how the internet works saying big words like "traceroute" and "backbone". Here's a little experiment: try tracerouting to a bunch of different places on the internet such as google, amazon, guildwarsguru.com, etc. See how on many of them traceroute just stops working after a few lines? Go a dozen hops away on the internet and traceroute and the like stop being reliable. The packets are being dropped because of strange TTLs (which can indicate all sorts of things including internet attacks), not because no routes exist.

If you want to do serious end-to-end bandwidth and latency measurements, traceroute doesn't do it. Traceroute only (kinda sorta) shows the route. You don't care about the route, because you know that you can connect to GW. What you care about is bandwidth and latency. use a tool like pathChirp or netest. If you have lower end-to-end bandwidth to guildwarsguru.com than to GW's servers, then that there is your problem.

Of course knowing where the problem exists is only half the battle. If somewhere on the path between your computer and GW's servers some service provider is playing silly QoS games, you're basically SOL.

Arkantos · Sep 03, 2006, 04:07 AM // 04:07

Quote:

Originally Posted by samifly

For all of you who arent suffering from lag, great, congratulations, you are one of the lucky few. Half of the community is suffering from this lag. This is not, a computer issue, if 4 people disconnect at the same time.

If half of the community is, then theres no lucky few. Theres a lucky half

TheProject · Sep 03, 2006, 04:13 AM // 04:13

ive had less lag as of late, but i have to download 2 ****ING MEGABYTES OF USELESS DATA EACH TIME I WANT TO CHANGE A CHARACTER OR GET BOOTED OFF!

EternalTempest · Sep 03, 2006, 04:18 AM // 04:18

Some stats. and NCSoft 1st response is update my video card driver... yeah sure thats the problem...

This was 9/1/06 around 1-3 am US / Eastern Time

[QUOTE]

Computing statistics for 15 seconds...
Source to Here This Node/Link
Hop RTT Lost/Sent = Pct Lost/Sent = Pct Address
0 (Removed for obvious reason)
0/ 4 = 0% |
1 1ms 0/ 4 = 0% 0/ 4 = 0% dslrouter (Removed for obvious reason)
0/ 4 = 0% |
2 40ms 0/ 4 = 0% 0/ 4 = 0% (Removed for obvious reason)
0/ 4 = 0% |
3 41ms 0/ 4 = 0% 0/ 4 = 0% so-0-1-2-0.CORE-RTR1.CHI01.verizon-gni.net [130.81.16.248]
0/ 4 = 0% |
4 41ms 0/ 4 = 0% 0/ 4 = 0% so-7-2-0-0.BB-RTR1.CHI01.verizon-gni.net [130.81.20.56]
0/ 4 = 0% |
5 41ms 0/ 4 = 0% 0/ 4 = 0% so-6-0-0-0.PEER-RTR1.CHI80.verizon-gni.net [130.81.16.11]
0/ 4 = 0% |
6 41ms 0/ 4 = 0% 0/ 4 = 0% ge1-14.fr1.ord.llnw.net [68.142.72.201]
0/ 4 = 0% |
7 41ms 0/ 4 = 0% 0/ 4 = 0% tge1-2.fr3.ord.llnw.net [69.28.171.69]
0/ 4 = 0% |
8 95ms 0/ 4 = 0% 0/ 4 = 0% tge2-1.fr3.dal.llnw.net [69.28.171.197]
0/ 4 = 0% |
9 106ms 0/ 4 = 0% 0/ 4 = 0% tge1-2.fr3.phx2.llnw.net [69.28.171.129]
0/ 4 = 0% |
10 106ms 0/ 4 = 0% 0/ 4 = 0% tge2-1.fr4.phx2.llnw.net [69.28.171.102]
0/ 4 = 0% |
11 106ms 0/ 4 = 0% 0/ 4 = 0% tge2-1.fr3.phx1.llnw.net [69.28.171.89]
0/ 4 = 0% |
12 107ms 0/ 4 = 0% 0/ 4 = 0% tge1-2.fr3.lax.llnw.net [69.28.171.82]
0/ 4 = 0% |
13 --- 4/ 4 =100% 4/ 4 =100% ncsoft.ge5-11.br01.lax05.pccwbtn.net [63.218.73.2]
0/ 4 = 0% |
14 --- 4/ 4 =100% 4/ 4 =100% 216-107-255-70.plaync.com [216.107.255.70]
0/ 4 = 0% |
15 106ms 0/ 4 = 0% 0/ 4 = 0% 216-107-246-36.plaync.com [216.107.246.36]

Bolded is the 100% packet loss aka the part where it's not talking.

seut · Sep 03, 2006, 04:43 AM // 04:43

Quote:

Originally Posted by Gaile Gray in response to my traceroute

Ok, does that or does that not show it's not Guild Wars, but instead is the provider or the backbone over which the data travels? Or the provider?

Hop12 where the timeouts starts is named "ncsoft.ge5-11.br01.lax05.pccwbtn.net". It's probably some kind of firewall/ dispatcher/whatever from NCSoft that blocks ICMP packets (ping, traceroute and other programs use that).
My traceroute shows, that my connection up to NCSoft's area of influence might be a little high in it's round-trip, because i play on US server from europe, but it clearly shows it's not a problem on my side or the way to the first technical equipment that carries the name NCSoft on it (i didn't post my gw.exe -diag log that shows no packetloss).
Also as i have stated before: the lag spikes/disconnects are experienced simultaneously by people on 4continents in different regions of the game (meaning different instances that are probably hosted on different servers of Anet's network). It's either a problem on Anet's doorstep, or with the servers.

my personal background: i'm a Diplom Informatiker(fh) (that's a degree somewhere between bachelor and masters in the field of computer science) and i've financed my studies by working in technical support of a major company... trust me: i eliminate problems on my side first, before i go and blame others.

gr3g · Sep 03, 2006, 05:11 AM // 05:11

Quote:

Originally Posted by seut

NCSoft that blocks ICMP packets (ping, traceroute and other programs use that)

Traceroute uses UDP, not ICMP. Or, to be even more precise, it sends UDP packets with low TTLs, and uses the RFC 792 ICMP Time Exceeded Message the router sends on packet drops at 0 TTL to calculate the route.

It could well be that the router you mention is dropping packets without sending a TEM (which is, technically, a violation of RFC 792), but it could be any number of other reasons why the packet was dropped. Traceroute will give you no clue as to the reason.

seut · Sep 03, 2006, 05:30 AM // 05:30

Quote:

Originally Posted by gr3g

Traceroute uses UDP, not ICMP. Or, to be even more precise, it sends UDP packets with low TTLs, and uses the RFC 792 ICMP Time Exceeded Message the router sends on packet drops at 0 TTL to calculate the route.

unfortunately Microsoft programmers don't read RFCs...
from http://en.wikipedia.org/wiki/Traceroute
The traceroute utility usually has an option to specify use of ICMP echo request (type 8) instead, as used by the Windows tracert utility.

gr3g · Sep 03, 2006, 05:35 AM // 05:35

Fine, but the end result is that traceroute is not giving you what you would like to know, which is end-to-end bandwidth and latency. For TCP mind, as that's what GW uses! For all we know TCP takes an entirely different route as UDP and ICMP.

seut · Sep 03, 2006, 08:21 AM // 08:21

My assumption that a traceroute would matter is based upon the commands that "gw.exe -diag" uses. It's using 'pathping' which is Microsoft's modern equivalent to traceroute.
But apart from our technical chatter: the recent rise in lag spikes and err7s affecting people in different locations (real life and gw-instances) simultaneously should be evidence enough that there is nothing wrong on the individual user's machines. It doesn't need to be Anet's fault, maybe it's the recent major update for Auto Assault that's clogging NCSoft's network pipes.
All i want from Anet is acknowledging that there might be a problem and that they look after it (what else is their Network News page for?).

Nexus Icon · Sep 03, 2006, 08:36 AM // 08:36

It's gradually becoming more and more apparent that the problem is neither at the user's end or ANet's end, but at some point in between.
However, it is up to ANet to fix it. You know, chuck money at the problem or whatever, just fix it.
Why?
Because if people keep having problems that show no signs of going away, they'll gradually stop playing, which means they'll stop buying newer content too.
After all, why buy something new for a game that is painful to play?

I know that I've stopped playing GW as much over the past week due to these problems, and if it continues for much longer I'll stop playing almost completely, only checking in at monthly intervals.

Evilsod · Sep 03, 2006, 12:46 PM // 12:46

So i guess its not Anets problem that 6 times last on Friday my entire team, and practically everyone else in HA had been getting booted off with code40s? Or that just by standing there you get lag or err7. This lag is getting beyond playable sometimes.

Gaile are you entirely sure your servers are quite what you make them out to be? Seems like you all seem to think they could handle the insane amount of people online for this weekend event.

seut · Sep 03, 2006, 01:03 PM // 13:03

My last lag spike is more than 24hours ago... *cross fingers*