[WebNoob]

Hello all,

This annoying issue is still cropping up for me (and apparently a number of people on the HLDS mailing list as well), but there never seems to be answers forthcoming - so I'd like to know if many people are still experiencing these sporadic server "hangs".

The problem:

Every now and then, a TF2 server will just "hang". It won't crash, as the server is still running, but it will disconnect all the players and remain "hung" (preventing players from joining) until it is manually restarted. Since I'm using TCAdmin across all of my servers, the server monitor does not restart the instance because it thinks it is still running, which means several times per day, server will get cleared out until I notice that they have hung again.

My server specs:

Windows server 2008R2
Dual quad core/12GB ram
Less than 1 32-slot TF2 server per core (currently only running 5 servers per DQC box)

My command line:

-game tf -console +ip 12.34.567.89 -port 27015 +maxplayers 32 -tickrate 66 +rcon_password dfhasjfhsjk +exec 1234576_server.cfg +map cp_dustbowl +tv_enable 0 -condebug -nocrashdialog -replay

What I've tried (and still experienced "hung" servers):

1. Running "vanilla" with no mods (clean install)
2. Running with and without replay
3. Running with just stock Sourcemod install

I know there are quite a few people in the same boat, and I'm surprised it hasn't been fixed yet (been going on for almost a year), but with 20+ TF2 servers, I'm getting 3-4 of these per day now, and it's driving me nuts.

Linux operators have suggested several methods to try and diagnose these things, but I haven't really heard or seen a reliable way to get to the bottom of the problem on a windows server.

Any advice on how to diagnose this and fix it once and for all would be greatly appreciated. If I need a 3rd party tool, so be it.

Note: The only way I've seen the issue first hand when I've been on a server is when half a dozen or more players leave a server all at once, it will sometimes hang.

[Snaggle]

Same problem for me too. It's a pain as it won't auto-restart itself as the process is still technically running.

[11530]

I've been getting these hangs myself, which is annoying since I don't want to have to keep an eye out for these problems and do a manual restart.

Soon I'm going to try something I found roaming the servers today: I wrote a post (#2) 20 mins ago elsewhere which can give more details: http://forums.alliedmods.net/showthread.php?t=183085

Fingers' crossed.

[Nomarky]

I've been looking for something to deal with this problem on Windows servers too. As I see it when the server hangs the process itself is still running, so it won't automatically restart because it hasn't actually crashed. The way to deal with this problem is to ping the server, then restart it if it fails to respond to ping requests after a set amount of time (say a minute or two) to avoid a restart being triggered by a map change or a short network timeout.

The closest thing I have found is a php script here but I couldn't figure out how to add back the retry attempts code, and then I forgot about it...

If someone could figure out how to put those retry checks back in then the script could be run as a scheduled task every 5 minutes or so and would pick up servers that have hung, minimising downtime.

[WebNoob]

Yep - two more of these today alone. I think it's great that people are trying to find ways to restart the servers automatically in the event of a "hang", but ideally I would rather find the cause and fix it at the source. This has been a problem for a significant amount of time now, and were it not happening, some of our servers would literally be full for days (our record before this issue cropped up was 13 days straight).

My guess is that we need some kind of monitoring/logging system that can spit out a verbose diagnostic log telling us exactly what is happening when the server hangs like this. The onyl way it is ever going to get fixed is if we can go to Valve and show them what is causing the hangs.

[ReFlexPoison]

Well I am almost certain it is faulty or old plugins that have been causing these problems.

Post your plugin info of what plugins you are running would probably help out the community to find the problems.

Vanilla installs shouldn't have these problems.

[Powerlord]

Is this just happening on Windows, or does it affect Linux as well?

[11530]

A better question would be: does anyone experience this problem with something other than TF2?
If not, feel free to all post your plugin/extension lists.

Originally Posted by Powerlord Is this just happening on Windows, or does it affect Linux as well?

Yea, happens to me as per my #3 post, and I use lightweight Archlinux.

[ReFlexPoison]

These problems of course could be just srcds having one of its thousands of problems.
If you install a vanilla server and this occurs then the above statement is probably true.

[Nomarky]

It has been noted on the mailing list that it happens with and without mods