OutOfMemory Exceptions, Native Stacktraces with gdl.

[Hawkzy]

Hey there.

So my server was okay for a good twenty four hours then it started throwing OutOfMemoryException(s), multiple of them. It seemed to crash after a good hour of the server being online, maybe sometimes more. Basically, it'd just be half way through the lava/water dropping and it'd just die, wouldn't restart itself. It'd just throw that exception and that was the end of that.
Also, this has only happened a couple of times but another error which resulting in a shutting down server was some thing called 'gdl' throwing a 'Native Stacktrace'. It'd just give me some gibberish with numbers, like 0x____ numbers, and it'd drop, and that was it.
Yeah, this'd do it all when it has about 26 to 32 people in the server.

Server: [HAWKZY] Lava Survival
Max Players: 32
Server Specs: Linux CentOS5 with Mono, 386MB RAM [762MB Burst]. No idea what the CPU is, sorry.
From: onedollarvps.com [The VPS 3 -I also have VPS 5 if you think it'd need more memory, I can always try running it on VPS 5].
MCD Version: 5.4.0.0

Any help on this would be greatly appreciated, thank you!
-Hawkzy

[Hawkzy]

I'm trying low lag mode now, because it seemed to work beforehand. I'm keeping a close eye on /serverreport and we just had a round of water where it went from 90MB usage to 160MB usage, but next round has loaded and it hasn't dropped. My thoughts are it isn't letting go of unused memory, so it's causing itself to throw the exception. In the meantime whilst I test lowlag mode, any help would still be greatly appreciated.
-Hawkzy

[Hawkzy]

Just took some numbers from a lava map, I will take ones from a water map next. Format below is RAM Usage, CPU%, (Players amount - current action)
161MB, 7% (22 players - 3 mins to lava)
161MB, 10% (23 players - 2 mins to lava)
161MB, 9% (24 players - 1 min to lava)
161MB, 7% (25 players - lava dropped)
162MB, 42% (24 players - 4 mins to go)
162MB, 84% (24 players - 3 mins to go)
162MB, 43% (24 players - 2 mins to go) [Map now convered in lava]
162MB, 7% (24 players - 1 min to go)
162MB, 6% (24 players - end of game)
162MB, 12% (26 players - map just swapped)

[Hawkzy]

228MB, 13% (32 players - 3 mins to water)
228MB, 18% (31 players - 2 mins to water)
228MB, 14% (31 players - 1 min to water)
228MB, 15% (31 players - water dropped)
30 seconds later, server crashed.

Yep, it's the active_cold_water. Changing that to ahl. Still got the Native Stacktraces to sort out, so next time it happens, I will report here.

[g0d01w4r99]

To try and see the error, I raised my max player limit to 35 players. The amount of players was constantly jumping between 30 and 35 after 2 hours of hosting. Haven't gotten any errors.
Cpu Specs:

Intel Core 2 Quad CPU Q6600 @ 2.40 GHz (4 CPU's)
8 GB of RAM DDR3. 2 GB per slot in 4 slots.
10MB of upload speed. (optic)
And a ASUS motherboard to support the above.

Only solution I see to this is that you increase your RAM capacity. 400MB of RAM is kinda a bad idea to host a lava server with.

I have never even seen the error, so I cant help you further, unless you post the full error log. I may be able to examine that. Hope this helped.
Cheers.

[Hawkzy]

In that case, could you, or anyone else please recommend a minimum amount of RAM I should have before running a lava survival?

[g0d01w4r99]

1 GB. I'd suggest 2 gigs though, other process's may screw things up.

[Hawkzy]

Ok, just bought a server with 680MB of RAM and 1.2GB of burst RAM. Also, that stacktrace error happened:

Code: Select all

mmap(...PROT_NONE...) failed
Stacktrace:

  at (wrapper managed-to-native) System.GC.InternalCollect (int) <0xffffffff>
  at System.GC.Collect () <0x00019>
  at MCDzienny.CmdGoto.Use (MCDzienny.Player,string) <0x0044b>
  at MCDzienny.LavaSystem.LavaThread () <0x0134b>
  at System.Threading.Thread.StartUnsafe () <0x00057>
  at (wrapper runtime-invoke) object.runtime_invoke_void__this__ (object,intptr,intptr,intptr) <0xffffffff>

Native stacktrace:

        mono [0x80e0c5b]
        /lib/libpthread.so.0 [0xa7efa0]
        /lib/libc.so.6(abort+0x101) [0x4eb701]
        mono [0x822b757]
        mono [0x8220037]
        mono [0x8229ed9]
        mono [0x82211c8]
        mono [0x82208c0]
        mono [0x822126c]
        mono [0x82212b7]
        [0x751eb4]
        [0x751e0a]
        [0x8d0644]
        [0x750de4]
        [0x791e18]
        [0x791cdf]
        mono [0x8062438]
        mono(mono_runtime_invoke+0x40) [0x8163920]
        mono(mono_runtime_delegate_invoke+0x36) [0x8163b76]
        mono [0x813ad62]
        mono [0x81faf45]
        mono [0x8224185]
        /lib/libpthread.so.0 [0xa76832]
        /lib/libc.so.6(clone+0x5e) [0x59345e]

Debug info from gdb:


=================================================================
Got a SIGABRT while executing native code. This usually indicates
a fatal error in the mono runtime or one of the native libraries
used by your application.
=================================================================


[dzienny]

This error is probably caused by the garbage collector malfunction. Make sure that you have the latest mono installed and when you run the application pass the --gc=sgen switch to the Mono. For more information read this: http://en.wikipedia.org/wiki/Mono_(software)#Garbage_collector .

[g0d01w4r99]

Ah, Mono problem, I never owned a mono, so I couldn't tell you, luckly, Dzienny has some experience with this