You could kick all the players from the server with the message
Shutting down, and then send SIGINT. You'd need to find a way to
flush the logs before sending SIGINT though.
On Nov 20, 2007 11:29 PM, Nye Liu [EMAIL PROTECTED] wrote:
On Tue, Nov 20, 2007 at 08:19:52PM -0800, Alfred Reynolds
Has anyone been able to get UDP logging to work in TF2? I've got it
set to send events on a port on localhost with logaddress_add, yet
the server never sends any UDP packets. It's logging to the console
and to file fine, and log says it's logging to UDP, yet my log
listener receives nothing, and
Have any of you had any memory usage problems with the TF2 dedicated
server? When I first start a 24 player server and fill it up with
players, top records srcds_i486 using around 8% of memory. However if
I leave the server running, memory usage slowly balloons. Here's some
output from top:
top -
Firstly: Who should I be sending bug reports to for the Team Fortress
2 dedicated server on Linux? Providing an entire core dump on a
mailing list is more than a little excessive.
Anyway, this is with the following version:
Protocol version 11
Exe version 1.0.0.5 (tf)
Exe build: 11:58:55 Oct 9
orangebox is a separate engine. cd into the orangebox directory and
run the srcds_run executable in that folder, along with -game tf.
On 9/26/07, Steve Johnson [EMAIL PROTECTED] wrote:
I'm trying to run a TF2 server on a CentOS 4.5 server with the following
command:
./srcds_i686 -secure -game
The lack of i686 and AMD-specific binaries in the Linux release is
interesting. Can you make any comment on this? Are optimized binaries
planned for release? And will the performance difference be
appreciable?
On 9/20/07, Jason Ruymen [EMAIL PROTECTED] wrote:
Updates to the Team Fortress 2
I think the more jarring effect that this creates is how the map
changes as soon as a player connects and joins a team. They spend all
that time connecting to the server, loading the map, and joining a
class or a team or what have you, and then they have to do it all over
again.
At least that's
If you've been running into issues with the server crashing, you might
find running the server under GDB useful for seeing where the server
crashed. Of course the only problem with this is when GDB stops running
the program, it just returns a command prompt and awaits your input,
instead of
Yeah, I'm aware, but as I said, the server won't even run under anything
but gdb. I had used ulimit and -debug in the past, but with this
particular crash the core dumps always turned up unusable. For some
reason it's working now though, so hopefully I won't have to use gdb if
this bug is fixed
If you're talking about the Source Dedicated Server, there aren't any
64-bit binaries released yet. I presume the engine_amd binary is just
optimized for AMD CPUs in general.
If 64-bit binaries were released, you'd need one for every binary in the
bin/ folder, as well as one for all the mods you
Well, I've never seen this happen before, but my server just stopped
running and printed this:
pipes.cpp (386) : Stalled cross-thread pipe
pipes.cpp (386) : Fatal assert failed: ../clientdll/pipes.cpp, line 386.
Application exiting.
Program exited with code 01.
Since that Stalled cross-thread
It's still segfaulting for me on my 2.4 kernel.
Alfred Reynolds wrote:
We have released a beta update for the Source and Half-Life 1 Linux
dedicated server. Run the hldsupdatetool with the -beta linux option
to get this update.
The update fixes the crash on startup on some Linux kernel
Run the server under gdb and try to reproduce the bug and post a
backtrace of the segfault here. If you're not sure how to run gdb, try
making a script like this:
#!/bin/bash
export LD_LIBRARY_PATH=.:bin:$LD_LIBRARY_PATH
screen -U -A -m -d -S css1 taskset 0x000C -- gdb --args \
./srcds_i686
Err, rather:
#!/bin/bash
export LD_LIBRARY_PATH=.:bin:$LD_LIBRARY_PATH
gdb --args ./srcds_i686 -debug -game cstrike -norestart +developer 1
Accidentally left in screen and taskset, which I use on my machine.
___
To unsubscribe, edit your list
http://i.dackz.net/misc/code/rcon.class.txt - that's what I use. It's
basically what was posted on the kquery wiki, but with some stupid bugs
fixed. I can't actually remember what I fixed, but it works fine now!
Alex Bergin wrote:
Just wondered if anyone knows of an example using PHP to connect
Does the Linux server console support multibyte characters though? I
have my terminal and server setup to display them, yet the server
console prints out characters for each byte.
___
To unsubscribe, edit your list preferences, or view the list
Actually I think I've got it displaying multibyte characters now, but
inputting them doesn't seem to work.
___
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
Disable SourceTV. The latest update has broken it.
___
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
As far as I can tell, sv_allowupload only applies to cheer sound files,
which most people don't even know about. You can try setting
net_maxfilesize to 0, but you'll still get that fragments from file crap
in your console.
I really wish there were a way to prevent them from uploading period and
This consistently happens around 2 minutes after the server is started
with SourceTV enabled and many people connected and playing. I'm not
running any server plugins and I'm not running sourcemm/metamod: source
(it doesn't even work!)
Here's a backtrace from gdb:
Program received signal
After some more testing, it seems to be the most pronounced on the
dod_flash map.
___
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
I definitely hope this is fixed soon. Doesn't sound like it'd be hard to
fix and it's pretty annoying, especially when physics objects are
actually solid.
Ben Daniels wrote:
has anyone else noticed that if you use the sv_Turbophysics 1 command, then spectators
can use objects like barrels and
22 matches
Mail list logo
