Minor Packet Loss Issue
 

I'm running a few servers on CentOS and a couple of the regular players are experiencing minor PL, myself included. Generally it hovers between 1-5 PL, which is almost unnoticeable in terms of gameplay, but still annoying. I was wondering if there were any setting in the ini or the server itself that could help improve this. Our tickrate is around 60, but seems to happen regardless. We have full access to the box.

Thanks.

What type of kernel are you using? what about the eth interface any virtuals?

Linux version 2.6.32-431.3.1.el6.x86_64 ([email protected]) (gcc version 4.4.7 20120313 (Red Hat 4.4.7-4) (GCC) ) #1 SMP Fri Jan 3 21:39:27 UTC 2014

eth0 Link encap:Ethernet HWaddr 00:25:90:A9:EE:66
inet addr:50.31.113.138 Bcast:50.31.113.139 Mask:255.255.255.252
inet6 addr: fe80::225:90ff:fea9:ee66/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:9000 Metric:1
RX packets:579599 errors:0 dropped:0 overruns:0 frame:0
TX packets:1143406 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:42619994 (40.6 MiB) TX bytes:58001972 (55.3 MiB)
Interrupt:16 Memory:dfa00000-dfa20000


AFAIK we aren't running any virtual servers/users whatever.

U r not alone. I have been working with gameservers for almost a month now, only seems to b effecting 2k4, not our 99 stuff, i have unfortunately tried almost everything i can think of, gameservers says its not on their end, something is defiantly going on, i only have 1 idea i havent tried, which is a problem with file(s), but if ur having the same issue i doubt it will fix it, but will know for sure in the next week at the most. I have had the server in 3 different locations, St.Louis, Chicago and NY/NJ, with 2 or 3 different boxes for each. I would like to ask u some simple questions like who r u running it from if not ur own box (if its CentOS i apologize in advance as i dont know what that is, altho my guess is an operating system of some kind lol), where is the servers locations and what gametype as well as style (normal weapons, instagib, ect.), maybe with both of us sharing info we can get this thing fixed

We have our own box because we got sick of shitty resellers. This is for the weapons CTF community. #2k4ctf irc.globalgamers.net . We have about 60 regulars.

CPU E3-1230v2
RAM 16GB DDR3
Operating System Centos 6 - 64 Bit
Raid None
Upgrade / Downgrade
Bandwidth 10 TB/Month on 100Mbps

also any time of day it happens particularly, how many ppl r on the server when it happens and how many maps r on the server? I had at least 1 other good question but forgot it lol, may just b easier to give me ur server ip address

Well a pug just started, you're welcome to spectate.

ut2004://50.31.113.138:7777
or
ut2004://50.31.113.138:1337

pw:pug

We never have more than 10, but usually 10.

We've actually deleted most of the non DM or CTF maps. Not that maps would affect the performance anyway.

ok ill check the setup in a bit, maybe find a common file or something, is this server in the US, if so what state?

It's in chicago. There is no way a common file is affecting it. I'm thinking it has something to do with the network setup.

It appears to b our game server company, so our problems have nothing related, sry, only thing i could tell ya is check ur ucc.log and fixed any errors in it, GL

How many IP addresses do you have? I'd suggst for protection reasons you have one IP as the root IP and gameservers on additionals. That way any ddos on the gameservers IP all your have to do is disable the virtual interface and your server remains accessible and you won't incur any traffic.

If you are with gameservers I'd suggest you ditch them and go dedi with a host that owns their network. Nlayer are very good. OVH are pretty good too and have a datacenter in USA.