We had a similar problem with random choke and lag spikes. Turned out to be due to having a large number of files in path.to.srcds/cstrike/downloads (mostly custom sprays) the servers was essentially hiccupping when parsing this directory prior to saving a new spray.
I now have a script that deletes the contents of this directory every night and restarts the server. May not be your problem but worth checking on. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of doomed999 Sent: Thursday, November 10, 2005 3:49 PM To: [email protected] Subject: [hlds_linux] Wierd Choke Problem Ok, I am seeing this on both hlds and srcds. I have not seen this problem before (on my servers) and isn't my rates doing it. The rates I have tried rate 30000 cl_update 100 cl_cmd 100 rate 25000 cl_update 80 cl_cmd 80 The best that seems to work is rate 30000 cl_update 50 cl_cmd 50 But there should be no reason I can't run 100 in my eyes. It normally spikes the worst when the bomb goes off and when the round restarts. With source it seems worse spiking up to the 30's with the 50 update rate. Server is running at 66 Tick. HLDS running at default and same thing. It is doing with with as little as one server on a machine. The machine specs are as follows Dual Xeon 3.0 2M cache 2G DDR2 533 ECC Running CentOS 4.1 Fully Burstable 100Mbps Port (Possible problem?) The only thing I can figure at this point is an overloaded network at the datacenter I am colocating through? I am at a complete loss. Please give your input. _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
