Do you happen to have a coredump? At 05:31 PM 10/3/2005, kama wrote:
I really getting frustrated with all these crashes I am getting. I've been trying to get any information on why these occur as it seems that I'm the only one that get those. This is what I have done so far. Tested on two hardware platform. FujitsuSiemens RX300 Dual 2.4GHz, HT enabled, 1GB RAM. HP DL380 G3 2.4, HT disabled, 1GB RAM. Both systems are running a recent FreeBSD 5.4-STABLE, but had the same problem when they where running 5.3-STABLE, which runned fine before the patch. linux_base-rh9, linux_base-suse5.2, linux_base-fedora-4. linprocfs is mounted. Fresh install of hlds without metamod or other plugins. tried with and without the -beta linuxbeta switch. tried with VAC2 enabled and disabled. tried with a single instance of hlds vmstats does not show any real problems. I have higher values on other machines that are comparable to those. It does not swap to disk. hlds is the only program that crashes. If it should be bad memory, other programs should occationally crash also. And it seems strange that memory should go bad on three different machines at the same time. The last line is most of the time: L 10/03/2005 - 19:54:43: "name<216><STEAM_0:1:23456789><>" STEAM USERID validated which makes me think its related to VAC2 thread. Even though I have started up the server without VAC2 it still threads the same way as it has it enabled. It started to happen after the latest patchset. It does not always create a coredumpfile, so I dont have anything real to go on. This happens at least 3 times per hlds instance, most of the time it average to 5 per instance. So I end up with 40-50 crashes per day in total. I getting out of ideas on what the problem is. I've been using FreeBSD and hlds for around five years now and have never had these problems. I have also other problems that seems to be network related. Could bad packets or burst of packets where the seq id get screwed up crash the server? Probably not, I would rather suspect the threading, ie the mainprocess tries to talk to a thread that are unavailable or not responding as fast as it should. I will later try a test running it on FreeBSD 6.0-RELENG. This is probably the last test I can think of. And I dont believe in any real success or improvement. /Bjorn -- // This is a comment. /* This is a comment. */ <!-- This is a comment --> # This is a comment. Have I provided enough comments? _______________________________________________ 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
