How can I reduce # of PID's used by server??

[Mr. Groch]

Hi!

I have a shell account where i can use only 9 PID's
and when I run tss2_rc2 server my account crashes...

Please - how can I reuce number of PID's used by server?
It's very important to me !! Help !!

[Brain]

What an odd shell acount... anyways, unless there's some undocumented switch that only the developers know: you can't.

[Randall_James]

I have 30+ servers, 200+ Channels only 1 PID

it was stated earlier but I have never heard of a way to limit the number of PID's. What about your web page? By the time you get a few viewers, Mysql, maybe a perl script or 2 running you would be in trouble......

[madcat]

hmmz how do you do that?

i have 10 (including the grep) so that makes 9

madcat:~# ps aux | grep -i server_linux | cut -d ' ' -f4 | sort | uniq | wc -l
10


but actually i'm also wondered why there are so many threads.
maybe linux.fan can help us out :)

[Cigar_Man]

I am also seeing a lot of PID's when starting the TS server on Fedora. I do not see this on a RH7.2 or a RH8 server.

Here is an example from a fresh server startup. I logged in remotely, set my screen session, then started the server with ./server_linux (note, this also happens using ./teamspeak2-server_startscript start). Here is what I see for this user:

normanb

Process ID CPU Started Command
1130 5.8 % 17:46 ./server_linux
1132 0.7 % 17:46 ./server_linux
1059 0.1 % 17:45 sshd: normanb@pts/0
1060 0.0 % 17:45 -bash
1097 0.0 % 17:46 SCREEN -A -m -d -S TS2
1098 0.0 % 17:46 /bin/bash
1131 0.0 % 17:46 ./server_linux
1133 0.0 % 17:46 ./server_linux
1134 0.0 % 17:46 ./server_linux
1135 0.0 % 17:46 ./server_linux
1136 0.0 % 17:46 ./server_linux
1137 0.0 % 17:46 ./server_linux
1138 0.0 % 17:46 ./server_linux

This is one TS2 server running the latest Version: v2.0.r20.b1 Linux. The user account has minimal rights and TS2 is running from his home directory.

Also, The TS2 server had been running for 35 hours and then it quit responding to client connections, although I could still login to the admin page at port 14534. There, it showed that the server was still running. The firewall showed the port open and netstat showed that port 8767 was being used by server_linux.

Is there a bug in this build?

Is it a bug while running TS2 on Fedora Core1?

I will keep digging!

[jean-luc]

Originally posted by Cigar_Man


Also, The TS2 server had been running for 35 hours and then it quit responding to client connections, although I could still login to the admin page at port 14534. There, it showed that the server was still running. The firewall showed the port open and netstat showed that port 8767 was being used by server_linux.



I had the same trouble this sunday...The server was running for 7 days and then quit responding to client..even thought th e admin page was available.

Might be a slight bug somewhere.

[risa2000]

I am running Gentoo, kernel 2.6.7-rc3 (was 2.6.6-rc3 too) and running teamspeak server produces 9 instance of "server_linux". Each instance allocating the same amount of memory (64MB of virtual space is quite a lot even for one instance, here it is x 9)
It looks like big wasting of resources to me. Is there any reason for this? Is it bug?
BTW "server_linux" is not a good name for teamspeak server, could anyone form devs change this?

[risa2000]

Just to add some more info:
Adding another server (from within superadmin mode) adds another 3 PIDS to the nine already running.
It looks to me, the main process should spawn working threads instead of forking the whole process. But I am coming from MS Windows world, so I may be wrong.

[guldi]

I am running Gentoo, kernel 2.6.7-rc3 (was 2.6.6-rc3 too) and running teamspeak server produces 9 instance of "server_linux". Each instance allocating the same amount of memory (64MB of virtual space is quite a lot even for one instance, here it is x 9)
It looks like big wasting of resources to me. Is there any reason for this? Is it bug?
If I remember right, the allocated memory is not for each instance. It's only for one, still 64 seems to be a lot to me ???
Again, I'm not sure but it reserves them based on the amount of users ??

could someone with more understanding confirm this ?

[Fazooly]

I have been having problems as well. Seems that there are multiple PIDs for each instance of TS I have running. Seems to run well, then all of a sudden clients can't connect. All "appears" to be running fine, but they can't utilize their TS.

We end up having to reboot to clear and reset everything.

I am not sure if it is a bug in the latest Redhat (we are currently running Enterprise) or a bug in TS, or a conflict between the two.

I hope the Devs look into this asap as it seems that it is becoming more widespread.

I look foward to a solution from TS.

[Peter]

First a note:

Seeing multiple PIDs is not a sign of a problem. The TeamSpeak server is a multi-threaded application, and linux displays threads very similar to processes. So, don't worry if you got 10 threads or something (or much more if you have many virtual servers).

If I remember right, the allocated memory is not for each instance. It's only for one, still 64 seems to be a lot to me ???
Again, I'm not sure but it reserves them based on the amount of users ??

could someone with more understanding confirm this ?

Threads share memory, so you definiatly should not add the usage of the individual threads together...The memory usage depends on many things...how many users are connected, if users are registered it takes even more...
Anyhow, TeamSpeak Server uses very little RAM, if you think its using too much, you are probably not reading the output right ;).

[Squillis]

I'm in FreeBSD and have 7 server_linux's running at 16MB a piece. This seems like a lot of ram for a server with 2 users logged in. So, is this behavior normal?

[Farlin]

I have a thread open on similar issue - http://www.teamspeak.org/forums/showthread.php?t=12916. Using 'top' I can see a lot of memory being used. The TS processes use more RAM than any other running process and additionally, it has multiple instances of the same process and each DO consume more memory. I have added the process totals together and I get about 2.1MB RAM per instance with approximately 9 instances running. That is 18.9MB RAM with no connections. As time goes on, the memory usage goes up to over 12MB RAM per instance, which is over 108MB RAM total, and eventually the server crashes to a <defunct> state. Seems like a lot of memory to me and I am reading the output right, I check the numbers and they do add up correctly.