Really Bad Lag

Ossified

Greetings,

So recently I've been having really bad lag, but ONLY when I'm either respawning or I join a team on a server that has a game running.

Every time I die and begin to respawn, I redplug for ~2 mins and eventually timeout from the server. Every time I join a team with a game already going, same thing happens. I redplug for a few minutes and then I just time out.

Anybody else experience this?

I've deleted my Workshop folder a MILLION times, I have 0 active/subscribed mods, and this is consistent across numerous servers.

Also, it's not just my general internet connection. I have great ping to these servers, it just happens when I'm respawning which I don't understand.

Any help/advice would be greatly appreciated!

IronHorse

It very well could be just your general internet connection.

While you are redplugged for ~2 minutes, immediately alt tab (run in windowed mode temporarily if you have to) and run Dos prompt.
Start button > cmd.exe > type "tracert www.google.com" without the quotation marks
Observe the results, post them here.

Spark utilizes a lot of bandwidth and sometimes it's too much for sensitive hardware like some routers. (often as a security measure)
Sometimes there are configurable options in their config panel related to flooding or overflow etc

ATF

I'm with Ironwall here, this sounds like an odd one.

The times you describe it happens are indeed times when a lot of updates are spontaneously sent to your client.
Try finding other times by left-clicking (switching spectate targets) when you're dead. It should send similar traffic spikeys. Or spectate in top down view and click about the map. That causes massive spikeys.

We're seeing people time out on our server all the time since 267, but those were reported to be actual client crashes (supposedly fixed in 269).
Your client isn't crashing. It even gives you the "Timeout" message and goes back to main menu?
Could you open your console and check your log? If it says something interesting, post it here (you can find the log file under " %appdata%\Natural Selection 2\log.txt " <- press windowskey + r and paste this)

Also try the command net_stats in the console. You may be able to find out whether the send or receive side gives in first when it happens. You can compare it to how it normally looks. If you see something noteworthy, take a screenshot (F12) and post it.

Ossified

Here's the output of a tracert to google during massive lag on respawn and eventual timeout from server:

Tracing route to www.google.com [74.125.228.208]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms 192.168.2.1
2 32 ms 25 ms 184 ms cpe-107-015-000-001.nc.res.rr.com [107.15.0.1]
3 13 ms 11 ms 13 ms 66.26.45.177
4 17 ms 19 ms 19 ms ae19.chrlncpop-rtr1.southeast.rr.com [24.93.64.2
]
5 24 ms 23 ms 23 ms bu-ether44.atlngamq46w-bcr00.tbone.rr.com [107.1
4.19.46]
6 23 ms 21 ms 21 ms 107.14.19.11
7 44 ms 49 ms 40 ms 72.14.219.56
8 30 ms 22 ms 21 ms 72.14.233.54
9 23 ms 24 ms 24 ms 66.249.94.6
10 27 ms 27 ms 28 ms 72.14.238.75
11 29 ms 28 ms 28 ms 66.249.95.228
12 31 ms 28 ms 28 ms 72.14.236.149
13 29 ms 37 ms 28 ms 72.14.233.91
14 30 ms 28 ms 28 ms iad23s23-in-f16.1e100.net [74.125.228.208]

Trace complete.


Nothing significant in console during lag spikes.

Only thing interesting/abnormal in net_stats is during the like spikes, my Choke goes to 100%.

Thanks for the help guys!

Ossified

Update:

I checked my router's security logs, and it is PACKED with "UDP Flooding to Host" and "UDP Flood Stopped". So I am assuming that NS2 is sending a lot of data at the same time, and my router's firewall thinks it may or may not be a DDOS attack, and is killing the packets.

I decided to put my computer into the DMZ of my router, so hopefully that helps. Will update thread soon after testing.

Ossified

Fixed! The problem was UDP packets flooding my router.

/Thread

IronHorse

Yea.. this is the oldest issue since launch of NS2 and has actually been somewhat reduced since then
(first weeks were LOADED with connectivity issues due to certain hardware flooding )

Glad its fixed.