Hi.
Recently we've had serious problems with atleast one of our Linux CSS servers. The server in question runs two 100tick gameservers, each 20 slots, one tends to be empty-half full and the other runs mostly 100% full. The servers run on ports 21000 and 21100.
In the past the server has run perfectly with zero lag and smoothly on 100tick.
Now it's started to run on 100% CPU load and massive lag and tickdrops follows. Today it got to 100% cpu load on 30/40 gameslots (20/20 and 10/20). Memory usage was low, ca 90mb for the most active server, stable. At the time we checked memory the other server had a mapchange and used about 160mb.
What we've tried:
- Full cold reboot daily (crontabbed) to get rid of any memory leakage, didnt work.
- Reboot of gameserver when load got that high, didnt work for long.
- Removal of secondary server, running only the most active server, resulted in 70% cpu load.
Server specs:
CPU: Dual Xeon 2.8GHz
RAM: 1 GB
Connection: 1 Gbit/s
OS: Linux (not sure which distribution, I'm unsure as I am not the server tech)
Command to start the server:
srcds_run -console -norestart -game cstrike -port 21000 +ip xx.xx.xx.xx -nocrashdialog -tickrate 100 +fps_max 600 +maxplayers 20 +map cs_office (Removed cfg-filepath for port settings and filepath for mani settings)
We hope somebody can shed some light on what could be wrong with our servers, it's frustrating as they've run perfectly in the past.
With regards
gnoble
Recently we've had serious problems with atleast one of our Linux CSS servers. The server in question runs two 100tick gameservers, each 20 slots, one tends to be empty-half full and the other runs mostly 100% full. The servers run on ports 21000 and 21100.
In the past the server has run perfectly with zero lag and smoothly on 100tick.
Now it's started to run on 100% CPU load and massive lag and tickdrops follows. Today it got to 100% cpu load on 30/40 gameslots (20/20 and 10/20). Memory usage was low, ca 90mb for the most active server, stable. At the time we checked memory the other server had a mapchange and used about 160mb.
What we've tried:
- Full cold reboot daily (crontabbed) to get rid of any memory leakage, didnt work.
- Reboot of gameserver when load got that high, didnt work for long.
- Removal of secondary server, running only the most active server, resulted in 70% cpu load.
Server specs:
CPU: Dual Xeon 2.8GHz
RAM: 1 GB
Connection: 1 Gbit/s
OS: Linux (not sure which distribution, I'm unsure as I am not the server tech)
Command to start the server:
srcds_run -console -norestart -game cstrike -port 21000 +ip xx.xx.xx.xx -nocrashdialog -tickrate 100 +fps_max 600 +maxplayers 20 +map cs_office (Removed cfg-filepath for port settings and filepath for mani settings)
We hope somebody can shed some light on what could be wrong with our servers, it's frustrating as they've run perfectly in the past.
With regards
gnoble