Hey, Yep, same here. Something clearly very buggy in srcds as that error is glibc (the system's C library) detecting memory corruption of some sort going on.
In fact I tried running it under valgrind to see if I could spot what was going on, but I gave up looking after over 100000 errors and a segfault. What seems to "fix" it for me is to set the following environment variable - basically disabling glibc's memory checks: export MALLOC_CHECK_=0 We can but hope that one day Valve will fix their engine. Cheers, Chris On 9 Oct 2010, at 21:38, Joonas Lehtolahti wrote: > I just tried and the same happened; crash right after trying to exec > server.cfg when loading the second map of The Sacrifice campaign. Although in > my case the crash cause was different, but originating from the same function > in tier0: > > *** glibc detected *** ./srcds_i486: double free or corruption (!prev): > 0x093aa308 *** > ======= Backtrace: ========= > /lib32/libc.so.6(+0x6c231)[0xf7618231] > /lib32/libc.so.6(+0x6dab8)[0xf7619ab8] > /lib32/libc.so.6(cfree+0x6d)[0xf761cb9d] > bin/tier0_i486.so(_ZN12CStdMemAlloc4FreeEPv+0x20)[0xf75648a4] > > (might be just different libc version handling free() differently causing the > crashing point to be different) > > Also there were some anomalies in the game itself. When there was supposed to > be tank fight by releasing the train door, there was only a horde of > infected, but no tank. When I tried to open the second door, it wouldn't let > me with message "Exterminate the tank!"... WHAT TANK? > Well, after the horde was eliminated to the last zombie, the game let the > door to be opened. Another weird thing was that after the pile of bricks > there were no infected at all. No specials or even common infected; the train > station was quite empty and so an easy run to the safe room, after which we > had the server crash. > > I also had those few harmless-looking error messages on loading the first > map, btw. There was also this: > > Connection to Steam servers successful. > VAC secure mode is activated. > NextBot tickrate changed from 0 (0.000ms) to 3 (0.100ms) > Scene '.vcd' missing! > Entity 202 (class 'infected') reported ENTITY_CHANGE_NONE but > 'm_clientLookatTarget' changed. > > Waiting for an update to fix this... > > > On Sat, 09 Oct 2010 15:23:46 +0300, ics <[email protected]> wrote: > >> As the topic says, when the sacrifice campaign changes from map 1 to map 2, >> the dedicated server always crashes. With plugins and without any plugins. I >> already mentioned this couple of days ago but nothing has happened since. >> There was an invisible L4D update today but even after that the crash >> occurs. Can you please fix this? Every other campaign is fine and servers >> run ok. Only The Sacrifice causes the crashes. There are 3 basic errors >> which are seen on map 1 load. Perhaps these lead to crash while loadint the >> 2nd map? Left 4 Dead 2 version and servers are fine. Only Left 4 Dead 1 is >> the problem. I get these on server console at map 1 load when players >> connect through the lobby. >> >> *** glibc detected *** ./srcds_i486: free(): invalid next size (fast): >> 0x0b93ba60 *** >> ======= Backtrace: ========= >> /lib/i686/cmov/libc.so.6[0xf7d9b1d4] >> /lib/i686/cmov/libc.so.6(cfree+0x96)[0xf7d9d186] >> bin/tier0_i486.so(_ZN12CStdMemAlloc4FreeEPv+0x20)[0xf7ce48a4] > > _______________________________________________ > To unsubscribe, edit your list preferences, or view the list archives, please > visit: > http://list.valvesoftware.com/mailman/listinfo/hlds_linux -- Chris Boot [email protected] _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
