Lag in Chicago (VDS)?
- Powerlord
- New to forums
- Posts: 13
- https://www.youtube.com/channel/UC40BgXanDqOYoVCYFDSTfHA
- Joined: Fri Dec 23, 2011 1:27 pm
- Location: cp_degrootkeep
- Contact:
Lag in Chicago (VDS)?
I'm currently leasing a dual-core VDS for my TF2 servers.
Some of my users are reporting high lag connecting to my servers. Also, the other day, all players got the disconnection warning (which you get if a server goes down) for nearly 15 seconds.
While I can't say the two are related, high lag is a particular concern of mine.
Are any other server admins in Chicago seeing this?
Some of my users are reporting high lag connecting to my servers. Also, the other day, all players got the disconnection warning (which you get if a server goes down) for nearly 15 seconds.
While I can't say the two are related, high lag is a particular concern of mine.
Are any other server admins in Chicago seeing this?
Admin, RUGC Midwest
Author of the MapChooser Extended plugin and a bunch of others.
Author of the MapChooser Extended plugin and a bunch of others.
Re: Lag in Chicago (VDS)?
What does the CPU/ram usage on your VDS look like? Did you make any recent changes to your server?
@Kraze^NFo> Juski has a very valid point
@Juski> Got my new signature, thanks!
@Kraze^NFo> Out of context!
@Juski> Doesn't matter!
@Juski> You said I had a valid point! You can't take it back now! It's out there!
@Juski> Got my new signature, thanks!
@Kraze^NFo> Out of context!
@Juski> Doesn't matter!
@Juski> You said I had a valid point! You can't take it back now! It's out there!
Re: Lag in Chicago (VDS)?
We haven't had any other complaints of this overall.
One other thing that commonly comes up with TF2 is the net_splitpacket_maxrate variable, which needs to be set high (such as 100000) or problems will occur. Here are some other things to investigate:
* Look at High CPU or memory usage, as kraze suggested. Have the task manager set to fast refreshes to spot any short spikes.
* Try without plugins and with our default server.cfg.
* Double-check that your servers are running with "high priority".
* Make sure you test with nothing else running (such as no webserver, mail server, etc), and try without an active connection through Remote Desktop or VNC.
* Make sure that the "server" Windows service is disabled, and try running with terminal services disabled or moved to a non-default ports (we see a lot of external IPs trying to guess passwords through it, which leads to performance problems for lower-end machines).
* Take a look at the net_graph 4 output and look for gaps, choke, packet loss, or server FPS dips, which will tell us more about what might be going on.
One other thing that commonly comes up with TF2 is the net_splitpacket_maxrate variable, which needs to be set high (such as 100000) or problems will occur. Here are some other things to investigate:
* Look at High CPU or memory usage, as kraze suggested. Have the task manager set to fast refreshes to spot any short spikes.
* Try without plugins and with our default server.cfg.
* Double-check that your servers are running with "high priority".
* Make sure you test with nothing else running (such as no webserver, mail server, etc), and try without an active connection through Remote Desktop or VNC.
* Make sure that the "server" Windows service is disabled, and try running with terminal services disabled or moved to a non-default ports (we see a lot of external IPs trying to guess passwords through it, which leads to performance problems for lower-end machines).
* Take a look at the net_graph 4 output and look for gaps, choke, packet loss, or server FPS dips, which will tell us more about what might be going on.
-
- Compulsive poster
- Posts: 60
- Joined: Tue Mar 29, 2011 7:42 am
Re: Lag in Chicago (VDS)?
This is more of a statement than a question.
I have also had some issues the past week or so on my 3 core managed VDS with 2 TF2 Servers (not chicago but atlanta). Not sure if it is related to any of the recent TF2 updates or the fact that it is holiday break for a bunch of people and servers are more active or if it is replay related.
Even back when TF2 first went F2P and both of my servers were full 24/7, they together only reached around 42% CPU usage at the absolute max, now I've seen it over 50% a number of times with the more intense maps. One full server by itself (with the other one empty) sits at about 25% CPU usage. That is about 5-10% higher than it used to be.
Here are the events that occured that never happened prior to the past week or so:
12/20 - High Memory Usage Auto Restart
12/23 - Automatic Remote Restart
12/24 - three or four mysterious server-wide "timeouts" where 20-23 people get Timed Out but the server stayed up.
12/25 - High CPU Usage Auto Restart
12/28 - one confirmed msysterious server-wide "timeouts" where 22 people got Timed Out
I disabled a bunch of plugins the other day but it didn't seem to help. I'm going to keep my eye on this and see if valve has an update or I simply have to upgrade packages to stop the timeouts/auto restarts.
I have also had some issues the past week or so on my 3 core managed VDS with 2 TF2 Servers (not chicago but atlanta). Not sure if it is related to any of the recent TF2 updates or the fact that it is holiday break for a bunch of people and servers are more active or if it is replay related.
Even back when TF2 first went F2P and both of my servers were full 24/7, they together only reached around 42% CPU usage at the absolute max, now I've seen it over 50% a number of times with the more intense maps. One full server by itself (with the other one empty) sits at about 25% CPU usage. That is about 5-10% higher than it used to be.
Here are the events that occured that never happened prior to the past week or so:
12/20 - High Memory Usage Auto Restart
12/23 - Automatic Remote Restart
12/24 - three or four mysterious server-wide "timeouts" where 20-23 people get Timed Out but the server stayed up.
12/25 - High CPU Usage Auto Restart
12/28 - one confirmed msysterious server-wide "timeouts" where 22 people got Timed Out
I disabled a bunch of plugins the other day but it didn't seem to help. I'm going to keep my eye on this and see if valve has an update or I simply have to upgrade packages to stop the timeouts/auto restarts.
Re: Lag in Chicago (VDS)?
untamed316, those high usage restarts do indicate that the server in question was using an entire processor core. If you continue to see these, I recommend turning off that feature through the "Managed panel" page.
I would also recommend making sure that the server is being restarted daily, as CPU and memory leaks are common.
I would also recommend making sure that the server is being restarted daily, as CPU and memory leaks are common.
-
- Compulsive poster
- Posts: 60
- Joined: Tue Mar 29, 2011 7:42 am
Re: Lag in Chicago (VDS)?
Well I unchecked the "restart if 100%" thing but it worries me that it actually has reached 100% of a single core. I experienced had another Auto-Remote restart tonight and Auto-Remote Repair. It appears to only be one of my servers having these problems. Both of my servers are set up exactly the same with the plugins, yet Server 1 (the most frequently full). I'll continue tweaking it.
Re: Lag in Chicago (VDS)?
I've also seen a few complaints on the mailing list recently about servers that stop responding and use 100% CPU, so it's very possible that a new game bug is the culprit.
Re: Lag in Chicago (VDS)?
It is because NFOServers in Chicago are garbage. Trust me, I finally had to switch hosts and problem was solves quickly with the new host. They blame us and these game architectures, which I can't replicate in a proper host.
Re: Lag in Chicago (VDS)?
stickmenz, stick to your flame in the flames section. Here, it is spam, as it does not apply to this customer's situation in any way.
His problem does appear to have been due to the specific TF2 game update.
His problem does appear to have been due to the specific TF2 game update.
- Powerlord
- New to forums
- Posts: 13
- Joined: Fri Dec 23, 2011 1:27 pm
- Location: cp_degrootkeep
- Contact:
Re: Lag in Chicago (VDS)?
I just realized that I never responded to this thread. I had ended up moving my NFO VDS server to the Texas (Huston?) location, where it remained operational until it was canceled earlier this month when its subscription ran out (I had left my old gaming group nearly 3 months prior and they moved to a different server).
My new gaming group already leases a Managed server from NFO.
My new gaming group already leases a Managed server from NFO.
Admin, RUGC Midwest
Author of the MapChooser Extended plugin and a bunch of others.
Author of the MapChooser Extended plugin and a bunch of others.