Constant 90002 Error

[Raist]

is it likely that SE will respond to so many people having 90k issues. The solution may be a slight change to our routers, but I see no official document on how to troubleshoot error codes on our end. I'd be happy to try things with some guidance.

help us SE, you're our only hope

It's been laid out many times in these connectivity threads...by myself and others. This post for Blizzard about testing the route to their servers is often referenced as well:

Running a Trace Route

It isn't usually so much an issue with your router, but more issues with how your ISP and/or their routing partners (Level3, Cogent, TATA, TiNet/SPA/GTT, or Verizon Business/ALTER.net) in order to get you to Ormuco (SE's ISP). Granted, there are times when it is something with SE's network or server configs--but the majority of the time we are finding signs of over-congested or flat out stalled hops in the routing along the way.

This is why using a VPN has been so successful for so many---it allows you to adjust the parameters of your routing to try to find a cleaner path to Ormuco.

Edit:
Just for a simple demonstration, here is my route tonight through my ISP (keep in mind, I've been working with them on cleaning things up... it used to be a nightmare before):

C:\Windows\System32>tracert neolobby02.ffxiv.com

Tracing route to neolobby02.ffxiv.com [199.91.189.74]
over a maximum of 30 hops:

  1     2 ms    <1 ms     1 ms  LPTSRV [10.10.100.1]
  2    35 ms    27 ms    47 ms  cpe-75-176-160-1.sc.res.rr.com [75.176.160.1]
  3    26 ms    30 ms    30 ms  cpe-024-031-198-005.sc.res.rr.com [24.31.198.5]
  4    15 ms    15 ms    14 ms  clmasoutheastmyr-rtr2.sc.rr.com [24.31.196.210]
  5    27 ms    26 ms    26 ms  be33.drhmncev01r.southeast.rr.com [24.93.64.180]
  6    33 ms    36 ms    33 ms  bu-ether35.asbnva1611w-bcr00.tbone.rr.com[107.14.19.42]
  7    29 ms    42 ms    31 ms  0.ae2.pr1.dca10.tbone.rr.com [107.14.17.204]
  8    54 ms    47 ms    46 ms  ix-17-0.tcore2.AEQ-Ashburn.as6453.net [216.6.87.149]
  9    89 ms    92 ms    88 ms  if-2-2.tcore1.AEQ-Ashburn.as6453.net [216.6.87.2]
 10    92 ms    93 ms    91 ms  64.86.85.1
 11    75 ms    75 ms    80 ms  if-10-2.tcore1.TTT-Toronto.as6453.net [64.86.32.33]
 12    90 ms    97 ms    98 ms  if-9-9.tcore1.TNK-Toronto.as6453.net [64.86.33.25]
 13    94 ms    92 ms    93 ms  if-7-2.tcore1.W6C-Montreal.as6453.net [66.198.96.61]
 14    91 ms    91 ms    88 ms  66.198.96.50
 15    89 ms    96 ms    92 ms  192.34.76.2
 16    87 ms    92 ms    94 ms  199.91.189.234
 17    94 ms    96 ms    93 ms  199.91.189.74

And here is a trace to the same server from the same laptop on my same wifi when using TunnelBear to their US location (looks like it is in the NY area):

Tracing route to neolobby02.ffxiv.com [199.91.189.74]
over a maximum of 30 hops:

  1    78 ms    78 ms    81 ms  172.18.11.1
  2    73 ms    73 ms    85 ms  107.170.26.254
  3    86 ms    77 ms    75 ms  192.241.164.233
  4    78 ms    78 ms    72 ms  66.110.96.21
  5   107 ms   108 ms   101 ms  63.243.128.121
  6   104 ms   106 ms   103 ms  if-5-5.tcore1.NYY-New-York.as6453.net [216.6.90.5]
  7   104 ms   107 ms   132 ms  if-11-2.tcore2.NYY-New-York.as6453.net [216.6.99.1]
  8   120 ms   107 ms   104 ms  if-12-6.tcore1.CT8-Chicago.as6453.net [216.6.99.46]
  9   109 ms     *      102 ms  if-22-2.tcore2.CT8-Chicago.as6453.net [64.86.79.1]
 10   101 ms   101 ms   107 ms  if-3-2.tcore1.W6C-Montreal.as6453.net [66.198.96.45]
 11   114 ms   114 ms   105 ms  66.198.96.50
 12    99 ms   105 ms    97 ms  192.34.76.2
 13   141 ms   161 ms   162 ms  199.91.189.234
 14   105 ms   105 ms   109 ms  199.91.189.74

Trace complete.

And this is when I tunnel to their Canada location (goes to OVH who has weird peering setup, and shoots me back to the US first--goes to show that sometimes the nearest VPN location is not always the best choice...and the same can be said for BGP routing and such the ISP's use):

Tracing route to neolobby02.ffxiv.com [199.91.189.74]
over a maximum of 30 hops:

  1    75 ms    84 ms    73 ms  172.18.12.1
  2   140 ms    75 ms    73 ms  192.99.8.252
  3    73 ms    74 ms    71 ms  bhs-g2-a9.qc.ca [198.27.73.95]
  4    80 ms    85 ms    88 ms  mtl-2-6k.qc.ca [198.27.73.6]
  5     *        *       88 ms  chi-2-6k.il.us [198.27.73.180]
  6     *        *        *     Request timed out.
  7   224 ms   228 ms    95 ms  if-22-2.tcore2.CT8-Chicago.as6453.net [64.86.79.1]
  8   100 ms    93 ms    91 ms  if-3-2.tcore1.W6C-Montreal.as6453.net [66.198.96.45]
  9   100 ms   101 ms   103 ms  66.198.96.50
 10   100 ms   104 ms   102 ms  192.34.76.2
 11   102 ms   100 ms   103 ms  199.91.189.234
 12    98 ms   101 ms    99 ms  199.91.189.74

Trace complete.

And last but not least is their tunnel to the UK (Netherlands, then London)...just for the helluvit:

Tracing route to neolobby02.ffxiv.com [199.91.189.74]
over a maximum of 30 hops:

  1   116 ms   121 ms   120 ms  172.18.12.1
  2   121 ms   118 ms   122 ms  46.101.0.254
  3   123 ms   121 ms   119 ms  5.101.111.229
  4   117 ms   119 ms   121 ms  xe-0-0-0-17.r02.londen01.uk.bb.gin.ntt.net [83.231.181.125]
  5   120 ms   122 ms   123 ms  ae-6.r02.londen03.uk.bb.gin.ntt.net [129.250.3.2]
  6   117 ms   122 ms   123 ms  ix-5-0.tcore1.LDN-London.as6453.net [195.219.83.185]
  7   220 ms   218 ms   209 ms  if-17-2.tcore1.L78-London.as6453.net [80.231.130.129]
  8   242 ms   215 ms   215 ms  if-2-2.tcore2.L78-London.as6453.net [80.231.131.1]
  9   215 ms   217 ms   217 ms  if-20-2.tcore2.NYY-New-York.as6453.net [216.6.99.13]
 10   218 ms   219 ms   215 ms  if-12-6.tcore1.CT8-Chicago.as6453.net [216.6.99.46]
 11   218 ms   214 ms   216 ms  if-22-2.tcore2.CT8-Chicago.as6453.net [64.86.79.1]
 12   220 ms   212 ms   218 ms  if-3-2.tcore1.W6C-Montreal.as6453.net [66.198.96.45]
 13   218 ms   220 ms   214 ms  66.198.96.50
 14   214 ms   214 ms   340 ms  192.34.76.2
 15   219 ms   218 ms   213 ms  199.91.189.234
 16   222 ms   217 ms   217 ms  199.91.189.74

Trace complete.

Each of those traces was from the same starting point to the same ending point, on the same system, same local router/modem combo--the only thing different was the use of a VPN to alter how things were routed. And, as you can see the results varied considerably based on how it was routed. For some odd reason, they all still ultimately used TATA (as6453) for the last peer towards Ormuco. Usually you'll see it change to Level3 or Cogent sometimes--which is sometimes what is needed, switching to an alternate peer because their exchanges are borked from higher traffic.

Notice how they all ultimately wind up at 66.198.96.50 and feed into 192.34.76.2 (the VPN ones also went to 66.198.96.45, but my ISP's route used 66.198.96.61). That is the hand-off from TATA into Ormuco (SE's ISP). From 4 different locations on two different continents, they ALL got shunted through the same third-party ISP and through the SAME exchange point. This happens often...and if there isn't enough bandwidth on either side of that exchange it creates an opportunity for a major bottleneck that can quickly spiral out of control and result in congestive failures. That isn't an election done by SE---that is the ISP's policies making those routing choices.

The same applies to your ISP. They decide not only how you are routed within their own network (which they can alter IF they wanted to), but also who you use to get from their network to SE's ISP's network---and they usually have multiple options on who that peering partner is. If they can't work out any issues with the current peer, they can switch you to another to see if it improves.