TS3 Server suffering from mass lost connections

Hi, since yesterday afternoon, I’ve got more than half of all my clients having connection lost and are unable to connect for A few minutes, sometimes hours.
This occurs instantly and there’s no 20 seconds delay between the lost and the disconnection.
We’ve been under attack before but usually the firewall handles it pretty well.

Update: It seem to affect other big servers, from what I’ve seen all inactive or muted/deaf users aren’t affected by any disconnection.
Also, we share in common the same server provider/reseller of: OVH.

Update 2: The issue comes from the firewall GAME that OVH uses. Once we’re under attack, it not only filters out the attackers but also drop legit traffic. It comes from a flaw in the said firewall.
Thanks to everyone for reaching out to me, lets hope OVH fixes the problem ASAP.

Update 3: They’re fixing it!!! Network & Infrastructure Status - AntiDDoS Game

you tried to restart vps or the server?

In what time you saw that ?
Actual time

We restarted multiple times.

Happened first time yesterday around 09:00 PM (21:00)
Today at ~01:00 AM, and multiple times during this afternoon.

What is time zone ?

I also saw something like that yesterday.

Well I’m really starting to think OVH changed something in there protection and now everyone is getting rejected from the port once someone attacks it.
But not sure about it.

Also, as I said, this happened even on a huge dedicated server that has the best protection out there.
We’re NOT the only one affected by this problem. This could also be an exploit someone discovered?

Why I don’t think it’s getting DDOSed is because there’s plenty of users that don’t get connection lost from it and they don’t have any lags at all.

1 Like

they may be getting a massive ddos attack , whe have same issue on our server and yep we at ovh,
its a good hosting but now have a little issue .
our server is works well no packet loss …etc but since 2 days some users are suddenly dropped and not allowed back but this is not a server error.
interesting becouse yesterday we restarted all things not just server, vps…etc… and its work until now.
is there a vps or firewall problem at ovh? :thinking:

1 Like

Monitoring the CPU, RAM and network usage, I don’t see anything abnormal exept it evidently decrease when this occurs.

1 Like

it doesn’t look too good

Hey,
Better to check through http://travaux.ovh.net/

It confirms that it comes from OVH.
Thanks !

1 Like

Edit: Still not fixed.

Uhm, but it related to SGP DC (Singapore).

Oh damn. You’re right. Didn’t read that one.

Really thought It was all over, I was blinded by my hopes :sweat_smile:

I give up for today, thanks anyway !

  1. Check pm
  2. Contact with OVH and send them link to this thread.

Hi, this happens again and again.
There’s no delay between the connection lost and the user’s activity to the server.
It’s instant, no 20 seconds delay before time out.
Here’s some logs:

> 22/04/2020 15:34:14	PktHandler	Info	Cleaning up connection because of 14 resends of COMMAND packet	
> 22/04/2020 15:34:14	PktHandler	Info	Dropping client 1323 because of resend timeout	
> 22/04/2020 15:34:14	VirtualServerBase	Info	client disconnected 'Crapulus'(id:51108) reason 'reasonmsg=connection lost'	
> 22/04/2020 15:34:14	PktHandler	Info	Cleaning up connection because of 11 resends of COMMAND packet	
> 22/04/2020 15:34:14	PktHandler	Info	Dropping client 1646 because of resend timeout	
> 22/04/2020 15:34:14	VirtualServerBase	Info	client disconnected 'Nezyum'(id:57226) reason 'reasonmsg=connection lost'	
> 22/04/2020 15:34:14	PktHandler	Info	Cleaning up connection because of 13 resends of COMMAND packet	
> 22/04/2020 15:34:14	PktHandler	Info	Dropping client 1454 because of resend timeout	
> 22/04/2020 15:34:14	PktHandler	Info	Cleaning up connection because of 10 resends of COMMAND packet	
> 22/04/2020 15:34:14	PktHandler	Info	Dropping client 1683 because of resend timeout	
> 22/04/2020 15:34:14	VirtualServerBase	Info	client disconnected 'tsukareta corporation ltd'(id:1871) reason 'reasonmsg=connection lost'	
> 22/04/2020 15:34:14	VirtualServerBase	Info	client disconnected 'Alexis Ross'(id:56501) reason 'reasonmsg=connection lost'
1 Like

Well there are a few attackers who have been launching massive attacks on medium/large teamspeak communities. I just hope you’re not the target.

Well I guess I am… our community is fairly big.
But what’s weird about it is that there’s no abnormal changes in CPU nor network usage.
I know we got attacked yesterday around 07:00 PM but that’s it.
Those drops and crash are occurring since 2 days every ~2 hours.