Random Disconnects From STO and Twitch

gt500dotorg

I've been having some odd disconnect issues lately where I get disconnected from STO, and at exactly the same moment OBS (Open Broadcaster Software) is no longer able to connect to Twitch.tv, which kills my stream.

When this starts happening, I open WinMTR and set it to accounts.crypticstudios.com which appears to resolve to 208.95.185.165, and while the issue is still happening I see the last two hops on the list are unreachable. I picked this hostname specifically because it is what TCPView said the game was trying to connect to (automatic hostname resolution, since all TCPView can see is the IP address). Here's my MTR log:

-- (Of course the forums wouldn't have a CODE tag, so here's a link to the log at Pastebin...)


The first unreachable address is perfectworldent-4.border11.bsn.pnap.net, which I cannot ping from the Command Prompt (although WinMTR can ping it) due to a hostname resolution error. WinMTR says the IP address is 216.52.61.78 when I double-click on the hostname, which most likely means that WinMTR found an alias for that IP address that my DNS isn't recognizing.

Regardless, I find it rather odd that my Twitch stream also disconnects when this happens, since I am streaming to the server in Chicago (or at least that's what OBS is set for). According to what I am seeing, the IP address 216.52.61.78 is for a server/router/switch/whatever in Boston.

Unfortunately I cannot get a proper traceroute or MTR log for the Twitch server I stream to, since it and an unknown number of servers in between it and me are not responding to ICMP requests (which basically means my WinMTR window is full of "No response from host" messages). The last hop on the list before that happens is xe-7-3-0.edge4.Chicago3.Level3.net which WinMTR says is at 4.53.98.45, and interestingly enough this server/router/switch/whatever is not on the list above from the traceroute on accounts.crypticstudios.com, so I am even more confused as to why both of them are dying at the exact same time.

Whenever it happens I start checking my Internet connection just to be sure that I can get out to other sites. Technically the MTR log proves that the Internet connection is still functioning since it can ping every server up to the switch/router/whatever the login server is connected to, but I still get paranoid when stuff stops working.

Now that I've written a blasted book for everyone to read, does anyone know what is going on? I know they had some downtime recently (last week?), and I think the issues developed after that.

captaingalaxy1

I have been having this problem too and its damm anyoying:(

qziqza

Hi there,

I have been experiencing alarmingly regular disconnects since last Thursday, post patch.
from my testing i would agree that the problem does seem to be somewhere along the
final few hops at the game end of things, this is the link to the thread i have reported on -

http://sto-forum.perfectworld.com/showthread.php?t=919211

tracert 208.95.186.32

Tracing route to 208.95.186.32 over a maximum of 30 hops

1 7 ms 7 ms 6 ms
2 9 ms 7 ms 5 ms virginmedia.net [62.252.174.45]
3 9 ms 12 ms 8 ms virginmedia.net [213.105.159.141]
4 9 ms 9 ms 9 ms virginmedia.net [62.254.42.113]
5 8 ms 9 ms 9 ms virginmedia.net [62.254.42.118]
6 30 ms 27 ms 27 ms virginmedia.net [62.253.188.158]
7 28 ms 29 ms 29 ms atlas.cogentco.com [130.117.15.5]
8 30 ms 29 ms 29 ms atlas.cogentco.com [154.54.74.89]
9 24 ms 26 ms 25 ms atlas.cogentco.com [154.54.62.145]
10 97 ms 97 ms 99 ms atlas.cogentco.com [154.54.44.237]
11 110 ms 98 ms 97 ms atlas.cogentco.com [66.28.4.42]
12 102 ms 107 ms 107 ms 38.111.40.114
13 98 ms 99 ms 97 ms 208.95.186.32

also took the liberty of running a nettest too..

contacting nettest server..
Local IP:
Ping: 85.7 msec
Port: 80: 548 KB/sec 13 KB/sec 582 KB/sec 500
Port: 80: 543 KB/sec 14 KB/sec 577 KB/sec 500
Port: 443: 580 KB/sec 13 KB/sec 616 KB/sec 500
Port: 443: 518 KB/sec 12 KB/sec 550 KB/sec 500
Port: 7255: 575 KB/sec 12 KB/sec 610 KB/sec 500
Port: 7255: 551 KB/sec 14 KB/sec 585 KB/sec 500
Port: 7003: 447 KB/sec 10 KB/sec 473 KB/sec 500
Port: 7003: 484 KB/sec 9 KB/sec 520 KB/sec 500
Port: 7202: 439 KB/sec 10 KB/sec 466 KB/sec 500
Port: 7202: 422 KB/sec 10 KB/sec 447 KB/sec 500
Port: 7499: 478 KB/sec 13 KB/sec 507 KB/sec 500
Port: 7499: 576 KB/sec 13 KB/sec 611 KB/sec 500
Port: 80: 471 KB/sec 10 KB/sec 499 KB/sec 500
Idle NIC bandwidth Send: 1 KB/sec Recv: 1 KB/sec

Regards

gt500dotorg

For some reason, no matter what server I select for my stream in OBS, when STO gets disconnected I also get disconnected from my Twitch stream. I've tried the servers in California, Chicago, New York, and Florida. In desperation I have selected the server in the UK this afternoon, but I haven't been disconnected yet, so I don't know if that helped or not.

ukcaptain420

is this problem related to this?

Where in the world are you?
Who is your ISP?

qziqza

ukcaptain420 wrote: »

is this problem related to this?

Where in the world are you?
Who is your ISP?

UK player here, and yup, i'm a VirginMedia customer too. can't fault them this time round though,
all else is running great with tracert and nettest pointing the finger of blame states side of things.

ukcaptain420

qziqza wrote: »

UK player here, and yup, i'm a VirginMedia customer too. can't fault them this time round though,
all else is running great with tracert and nettest pointing the finger of blame states side of things.

yeah that's the thing, i don't think it's the fault of sto, or virgin, just something along the way, i'm using EU proxy as work around and have been stable for days.

the OP isn't on VM, so either this problem is isolated to them or they are routing along some of the same path as VM

qziqza

ukcaptain420 wrote: »

yeah that's the thing, i don't think it's the fault of sto, or virgin, just something along the way, i'm using EU proxy as work around and have been stable for days.

the OP isn't on VM, so either this problem is isolated to them or they are routing along some of the same path as VM

even with the proxy im still struggling with stability, but it has lessened considerably since yesterday.. so fingers crossed. for a couple of days there, i was being dc'd pretty much every map switch, which when your doffing is very frequently. if i had a pound for every time i said 'thats it, no more, i quit.. logging back in' i would have very heavy pockets

gt500dotorg

ukcaptain420 wrote:

is this problem related to this?

It may be. I'm not entirely sure. No one else has run MTR, from the looks of it.

ukcaptain420 wrote:

Where in the world are you?
Who is your ISP?

In the real world I am in Indiana. In the game world it doesn't matter, the disconnects happen randomly.

My ISP doesn't matter, because the issue is clearly in Cryptic's datacenter according to MTR. When I see the last two hops in the traceroute (Cryptic's account server and the switch/router it is connected to) suddenly stop responding to pings the moment the disconnect happens, and every other hop along the route is fine, I know where the problem is.

I've been wondering if it was a DDoS attack aimed at Cryptic and Twitch (and perhaps other services I don't use), since both are going down at the same time. Unfortunately I cannot find out where the issue with Twitch is, because they have their stuff configured to not respond to ICMP requests. I can see all of the hops to Twitch are pinging fine, but none of the stuff in their datacenter will respond, regardless of whether or not my stream is working. The weird thing is that I have tried streaming to so many different Twitch servers in so many places, and had the same issue. That's why I suspect a DDoS, because it doesn't make sense for so many servers to go offline at the same time.

ukcaptain420

gt500dotorg wrote: »

It may be. I'm not entirely sure. No one else has run MTR, from the looks of it.




In the real world I am in Indiana. In the game world it doesn't matter, the disconnects happen randomly.

My ISP doesn't matter, because the issue is clearly in Cryptic's datacenter according to MTR. When I see the last two hops in the traceroute (Cryptic's account server and the switch/router it is connected to) suddenly stop responding to pings the moment the disconnect happens, and every other hop along the route is fine, I know where the problem is.

I've been wondering if it was a DDoS attack aimed at Cryptic and Twitch (and perhaps other services I don't use), since both are going down at the same time. Unfortunately I cannot find out where the issue with Twitch is, because they have their stuff configured to not respond to ICMP requests. I can see all of the hops to Twitch are pinging fine, but none of the stuff in their datacenter will respond, regardless of whether or not my stream is working. The weird thing is that I have tried streaming to so many different Twitch servers in so many places, and had the same issue. That's why I suspect a DDoS, because it doesn't make sense for so many servers to go offline at the same time.

Then my apologies for bothering you. I've said plenty that I am no expert, but I am a pretty decent problem solver. I still think your problems are linked to ours and I continue to disagree that this is cryptic issue.

Am I wrong? I honestly don't know. Yours could be an isolated incident with similar symptoms. It could be the same problem and be cryptics problem. It could be the same problem and be somewhere along the route(but your data shows otherwise). Some people have started to blame 'cogent' but I don't know enough about them or what they do(yet), to place blame on them either way.

All I know for sure is a large number of people are having the same problem, and the majority appear to be on the same ISP in the same country.

Does this mean the problem is isolated to them? No.

Does this mean it's the same problem? No.

Are the majority of STO players continuing to play without a problem? Yes.

This is an honest question(remember the no expert thing).

If there was a problem along the route, previous to the cryptic data center, that was OK when the MTR checked it, so marked it as pass, but then the problem occured while it was checking the data center node, where would it report the problem to be?