Whats the cause of Error 90002

[Saret]

I bought ARR 2 weeks ago, been playing for those 2 weeks without a problem on the ultors server, but since yesterday i get that error while playing (only while playing, no problem on the character select screen) after a random amount of time (could be 1 minute to 30 minutes) and i read in this forums that a lot of people have the same problem, all of them saying the same thing (myself included)

-their internect connection is fine
-it never happend before
-their connections nevers goes off for even a second
-Port forwarding doesnt work (then again i´ve seen so many different posts with different ports yet no official ports by SE)

So does anyone knows what exactly causes Error 90002? because i just dont know how i can fix this on my end (if it can be fixed on my end) if it started yesterday (september 19) and i dont have any problems with my Internet conection and there is no information on the support center about said error code.

EDIT: thanks Eye_gore, Raist and Risvertasashi, now i know why the error happens and got it fixed with WTfast.

[Eye_Gore]

Its a server disconnect error. Your equipment may be fine its somewhere along the route to the server that it gets a hiccup. Run a trace on you connection to their server. I don't know the servers address but someone here may. When you run the trace you'll see how many different "pipes" your connection is going through.

Get your command prompt up, type in tracert 199.91.189.36. See where that takes you. See if any timeout or have huge lag spikes (high ms). There are ways to get around the routing, I don't know how to do it, but again I'm sure someone here can point you in the right direction.

[Saret]

Its a server disconnect error. Your equipment may be fine its somewhere along the route to the server that it gets a hiccup. Run a trace on you connection to their server. I don't know the servers address but someone here may. When you run the trace you'll see how many different "pipes" your connection is going through.

Get your command prompt up, type in tracert 199.91.189.36. See where that takes you. See if any timeout or have huge lag spikes (high ms). There are ways to get around the routing, I don't know how to do it, but again I'm sure someone here can point you in the right direction.

I did the trace 2 times to be sure and this is what i got

Tracing route to 199.91.189.36 over a maximum of 30 hops

1 8 ms 8 ms 10 ms
2 * * * Request timed out.
3 9 ms 28 ms 8 ms ci2.0-1-1.cn2.gtdinternet.com [190.196.126.240]

4 8 ms 26 ms 8 ms 195.22.221.85
5 107 ms 117 ms 116 ms 195.22.199.179
6 111 ms 116 ms 111 ms 89.221.41.177
7 106 ms 110 ms 110 ms xe-11-0-0.edge2.miami1.Level3.net [63.209.150.16
5]
8 170 ms 173 ms 170 ms ae-11-11.car2.Montreal2.Level3.net [4.69.141.1]

9 328 ms 213 ms 221 ms ae-11-11.car2.Montreal2.Level3.net [4.69.141.1]

10 164 ms 168 ms 166 ms ORMUCO-COMM.car2.Montreal2.Level3.net [4.59.178.
74]
11 146 ms 149 ms 150 ms 192.34.76.10
12 150 ms 145 ms 145 ms 199.91.189.242
13 152 ms 151 ms 151 ms 199.91.189.36

Im getting a timeout on the second hop, but i have no idea of what to do with this information

[Eye_Gore]

That second one is your modem/router then to your IP. So it seems your equipment may be the issue. I'm not sure where to go with this either I'm not an IT guy, but maybe some one else will look at it and know. I know sometimes if UPnP is turned on it can be an issue, that would be a setting in your router. If you have a router to modem set up that second one would be the modem as the first is the router. Not that any of this really helps but atleast you know where the connection is being lost. That first one you may want to edit out thats your IP address not sure but it could be a security issue.

Another simple fix could be if you are running on wireless, you may be getting interference from something. Try a wired connection and see if that does anything. If it is wired, try a different wire.

[Saret]

From what i know a timeout from the router is not really bad as it is something they usually do because of priority, since it sending the package to another destination and only when the timeout continues then you have a conection issue, while in my case after the timout it continues to its destination, but then again thats what i learn from reading stuff of the internet, not sure if that timeout on the second hop which i guess is the router (i have a modem conected to a router and a cable conexion from the router to my pc) is actually messing with my conection to the server causing error 90002, since i had 0 problems from the past 2 weeks. I tried disabling UPnP but i still get that timeout on the second hop

[Raist]

You have too much jitter on the line, both locally and abroad (hops 3 and 4, then again at hop 9). Their shouldn't be extremely large variances between the response times at the same hop. This can be one of the signs of elevated congestion--any variances beyond a 50% difference can easily cause packets to arrive out of order, which may prompt a re-transmit. If things get far enough out of hand, it can cause a sort of throttling effect at the protocol level (TCP/IP is detecting congestion and tries to adjust the transmit windows), and if the delays cause things to get far enough out of synch the client/server rules may think you have lost connection and you loose your session.

Also note that the open world communications naturally refresh on a 300ms boundary, more frequently in instances or when you submit commands. Any response time approaching/exceeding 150ms+ will be prone to issues---one retransmittal can cause you to "skip a beat", so to speak. Note we aren't talking within a trace for that figure, but a ping response directly to the server. Tracert is for tracking patterns for signs of issues, not for measuring actual end-to-end response times. So, you may want to also run a ping test of about 20 or 30 pings as well with the -n switch (typically, if there is a troublesome delay in forwarding it will show within 50 pings...you shouldn't need a long sample to catch it). For example, to ping the neolobby02 lobby server (Aether) 25 times, the command would like like this:

ping neolobby02.ffxiv.com -n 25

Since some of the congestion appears to be closer to home, your ISP may be able to quickly remedy that part of it. But the more remote hop (the Level3 one) is third party. It is an ISP your ISP has entered into a peering agreement with, so they can report the matter to them for Level3 to address...but in all honesty, we've seen issues with their car2 hops in Montreal for a VERY long time. Talking back before XIV was even released (it's shown up in other complaints going back to 2010 and further back). So, it may be better for your ISP to route around them---they can and should have agreements with another top tier peering partner they can switch you to (Ormuco (SE's ISP) peers with 5 top tier ISP's in North America...big names like Cogent, TATA, and Verizon). The trick is getting them to do it. You will need to provide some sample data for them to work with and run tests with... hopefully you can catch some good samples to prod them into action.

Run resource monitor (resmon from search/run box) and get the IP address that FFXIV is actually using while you are fully connected and in the game by looking in the Network or TCP sections. Run your traces to that specific IP address. Run a few of them, trying to capture a sample when the connectivity is good and again when it is bad. Also take some against other services that appear good as well (youtube, netflix, another game, etc.) for a comparison. Save the results in a text file (notepad, wordpad, etc.) so you can forward them to your ISP's Tier3 support team. Then go to your ISP's support portal and open a ticket, including this data.

Here is a blog post from Blizzard about testing routes and such:

https://us.battle.net/support/en/art...e-and-pathping

It provides IP addresses for their servers. Can run traces to them for comparison if you need ideas.

Another thing you can try is to use a VPN. Most offer a free trial period or other limited free use plan that you can use just for testing. Try different tunnel locations, and if you can get one that is notably more stable, capture a trace through the service for comparison as well. Fixing the problem through a VPN pretty much proves there is a problem with how your ISP is routing you...so it is another very useful tool for fixing the problem, and not just a bandaid. Many have had to resort to using a VPN long-term because their ISP's aren't stepping up to the plate for them. Hopefully you will fair better if you can provide them with enough proof of smoke to prompt them to look for the fire.

If you are having trouble finding a good way to contact them online (some just give a number or generic email form), you can check the downdetector site to see if they have a listing for your ISP:

https://downdetector.com/companies
(Note they have links to other sites for different regions along the bottom of the page, for those not in North America)

They sometimes have links to forums or social media that you can use for requesting assistance. Even if you find a forum, it may be worth hitting their social media pages as well if you can. Don't try to dump the traces to Twitter or Facebook though, just a short note that you have captured delayed/lost packets and you need Tier3 support on the matter. With any luck, someone will reach out to you. If they don't have their own forum, you can check the DSL reports forum to see if they have one there:

http://www.dslreports.com/forums/all

[Saret]

I used Blizzard wow servers for comparison since i used to play wow and never had any problems and in both IP addreses everything works fine until hop 11 were i get a time out all the way to hop 30, but in their site it says that because of their server security you may get false latency numbers leading to request timed out messages, but as long as you reached their servers you should ignores those message and according to the tracer i did reached their servers in hop 8 in both cases.

i should mention (should have mention this from the very beggining since im guessing its really important information to understand my trace results) that I´m from chile and WoW has southamerican servers and north american servers, but the list doesnt make a difference between the two so im assuming those servers were located in NA since that post seems older than the release of SA servers, but numbers were stable in both cases.

i did a new tracing to ultors server today and the results were far better than yesterday, hop 3,4 and 9 being consistent this time (hop 3 and 4 stayed between 8 to 9 and hop 9 between 149 to 154, instead of that 300 jump, the rest was pretty much the same)

i also did a ping test to every single IP addres from the list (50 to every single one just to be sure) and i got a 0% loss in all of them so i went in and played the game and i still got error 90002 after 10 to 15 minutes, did a ping test again while the game was on, still 0%, so tried connecting the modem directly to my pc but it didnt make a difference, error keeps happening after a random amount of time

[Risvertasashi]

90002 means the server (or more likely an intermediate node) dropped your connection.

More than tracert, I recommend using winMTR (use the IP of your world server). This will show you if there's any significant lag or packetloss between yourself and the server. Note that the results will vary based on day and time of day, especially weekends, holidays, or other times of high saturation.

What you are likely to find is a Level3 or similarly disreputable provider's node between yourself and the server that is dropping packets. Use of a VPN can prevent this.

[Saret]

90002 means the server (or more likely an intermediate node) dropped your connection.

More than tracert, I recommend using winMTR (use the IP of your world server). This will show you if there's any significant lag or packetloss between yourself and the server. Note that the results will vary based on day and time of day, especially weekends, holidays, or other times of high saturation.

What you are likely to find is a Level3 or similarly disreputable provider's node between yourself and the server that is dropping packets. Use of a VPN can prevent this.

thanks for telling me about winMTR, its way easier and faster than doing cmd+tracer, and well you were right about level3

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 190.196.127.223 - 0 | 233 | 233 | 8 | 9 | 28 | 13 |
| No response from host - 100 | 47 | 0 | 0 | 0 | 0 | 0 |
| ci2.0-1-1.cn2.gtdinternet.com - 0 | 233 | 233 | 7 | 9 | 54 | 14 |
| 195.22.221.85 - 0 | 233 | 233 | 8 | 9 | 28 | 9 |
| 195.22.199.179 - 0 | 233 | 233 | 106 | 113 | 146 | 108 |
| 195.22.199.179 - 0 | 233 | 233 | 106 | 113 | 145 | 115 |
| xe-11-0-0.edge2.miami1.Level3.net - 3 | 213 | 208 | 105 | 116 | 217 | 106 |
| ae-11-11.car2.Montreal2.Level3.net - 0 | 233 | 233 | 163 | 208 | 419 | 187 |
| ae-11-11.car2.Montreal2.Level3.net - 0 | 233 | 233 | 164 | 207 | 396 | 186 |
| ORMUCO-COMM.car2.Montreal2.Level3.net - 0 | 233 | 233 | 169 | 189 | 208 | 196 |
| 192.34.76.10 - 0 | 233 | 233 | 144 | 150 | 213 | 144 |
| 199.91.189.242 - 0 | 233 | 233 | 145 | 149 | 196 | 145 |
| 199.91.189.36 - 0 | 233 | 233 | 146 | 152 | 170 | 146 |
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider

Hop 7 xe-11-0-0.edge2.miami1.Level3.net has a 3% packet loss.

So now what should i do know with this information? contanct my ISP? if so what should i tell them exactly?

[Risvertasashi]

You can contact your ISP (I certainly do) but it likely won't get anything done even months later. Level3 basically gives everyone the middle finger.

3% shouldn't be high enough to cause disconnects but that number might become quite a bit higher this evening and on the weekend. Keep an eye on it.

FWIW I use WTFast. Pingzapper, I hear, is also pretty good. Any VPN designed to fix routing issues should help, if this is in fact the problem. Keep an eye on it this evening and see where the numbers fall.