[asherkin]

Originally Posted by MaloModo Yeah lots of crashes since last 2 updates I think. I notice on https://crash.limetech.org/stats more than doubled at least short term and still above "previous averages".

That data is useless unless you understand what you're looking at (and drill down deeper than the main page) - there is absolutely no way to measure "crashiness" from the data on that page.

Surfacing the data for only labeled crashes would be a lot more useful for trend analysis - but requires manual analysis work to create the labeling rules.

[MaloModo]

Originally Posted by asherkin That data is useless unless you understand what you're looking at (and drill down deeper than the main page) - there is absolutely no way to measure "crashiness" from the data on that page. Surfacing the data for only labeled crashes would be a lot more useful for trend analysis - but requires manual analysis work to create the labeling rules.

I never said anything. I thought it was just interesting. But if "crashiness" goes up that is not anything? And sometimes correlation is at least in part causation. Just saying.

I believe in my particular case had to do around this whole assigning attributes amongst a few more things. For whatever reason things have become a lot more "unstable" with weird "problems". Even this whole thing around ' vs " in game was a pita.

[asherkin]

Originally Posted by MaloModo But if "crashiness" goes up that is not anything?

Crashiness going up is indeed important, but the graph on that page does not measure crashiness - just the rate that Throttle is processing dumps (most of the increases on the chart are from more servers using Accelerator, but relative stats are still meaningless since people only install Accelerator when they're getting crashes).

[MaloModo]

Originally Posted by asherkin Crashiness going up is indeed important, but the graph on that page does not measure crashiness - just the rate that Throttle is processing dumps (most of the increases on the chart are from more servers using Accelerator, but relative stats are still meaningless since people only install Accelerator when they're getting crashes).

Right..so "indirectly" that graph measures crashiness.


Anyway, I am not here to "argue". But I would like an answer about this admin_groups.cfg and admins.cfg posted. One other and and I have gone through the process and we don't see anything obvious. And I swear I have done this hundreds of times before.

[Powerlord]

Hmm... while this isn't necessarily major, I noticed that my server crashes on shutdown since the latest TF2 update.

Program terminated with signal SIGABRT, Aborted.
#0  0xf771d440 in __kernel_vsyscall ()
(gdb) bt
#0  0xf771d440 in __kernel_vsyscall ()
#1  0xf7457607 in __GI_raise (sig=sig@entry=6)
    at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
#2  0xf745aa33 in __GI_abort () at abort.c:89
#3  0xf7491e53 in __libc_message (do_abort=do_abort@entry=1,
    fmt=fmt@entry=0xf758eedc "*** Error in `%s': %s: 0x%s ***\n")
    at ../sysdeps/posix/libc_fatal.c:175
#4  0xf749c33a in malloc_printerr (action=<optimized out>,
    str=0xf758a9ed "free(): invalid pointer", ptr=0xb373918) at malloc.c:4996
#5  0xf749cfad in _int_free (av=0xf75d3420 <main_arena>, p=<optimized out>,
    have_lock=0) at malloc.c:3840
#6  0xf76799d9 in operator delete (ptr=0xb373918)
    at /valve/build/targets/src/gcc-4.6.1/libstdc++-v3/libsupc++/del_op.cc:47
#7  0xf7679a23 in operator delete[] (ptr=0xb373918)
    at /valve/build/targets/src/gcc-4.6.1/libstdc++-v3/libsupc++/del_opv.cc:32
#8  0xeb5e1b4c in PlayerManager::~PlayerManager() ()
   from /home/tf2/tf2/tf/addons/sourcemod/bin/sourcemod.2.tf2.so
#9  0xf745c543 in __cxa_finalize (d=0xeb62e200) at cxa_finalize.c:56
#10 0xeb5c3c63 in __do_global_dtors_aux ()
   from /home/tf2/tf2/tf/addons/sourcemod/bin/sourcemod.2.tf2.so
#11 0xf7731a5c in _dl_close_worker (map=map@entry=0xb312778) at dl-close.c:272
#12 0xf7732491 in _dl_close (_map=0xb312778) at dl-close.c:773
#13 0xf76c7dc2 in dlclose_doit (handle=0xb312778) at dlclose.c:35
#14 0xf772cc26 in _dl_catch_error (objname=0x90c1014, errstring=0x90c1018,
    mallocedp=0x90c1010, operate=0xf76c7da0 <dlclose_doit>, args=0xb312778)
    at dl-error.c:187
#15 0xf76c837c in _dlerror_run (
    operate=operate@entry=0xf76c7da0 <dlclose_doit>, args=args@entry=0xb312778)
    at dlerror.c:163
#16 0xf76c7df8 in __dlclose (handle=0xb312778) at dlclose.c:46
#17 0xf3301ea4 in UnloadInterface_MMS ()
   from /home/tf2/tf2/tf/addons/sourcemod/bin/sourcemod_mm_i486.so
#18 0x0b312778 in ?? ()
Backtrace stopped: previous frame inner to this frame (corrupt stack?)

It looks like SourceMod's player manager is trying to delete an invalid pointer.

[Wormy]

So Sourcemod may have been flawed after all, or is it too early to say?

[ocwoody]

Disabling all lunchbox items on my servers fixed my issue. Haven't received one crash since I've done so. I run both linux and windows servers.

[Dr_Knuckles]

Originally Posted by Powerlord Hmm... while this isn't necessarily major, I noticed that my server crashes on shutdown since the latest TF2 update. Program terminated with signal SIGABRT, Aborted. #0 0xf771d440 in __kernel_vsyscall () (gdb) bt #0 0xf771d440 in __kernel_vsyscall () #1 0xf7457607 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56 #2 0xf745aa33 in __GI_abort () at abort.c:89 #3 0xf7491e53 in __libc_message (do_abort=do_abort@entry=1, fmt=fmt@entry=0xf758eedc "*** Error in `%s': %s: 0x%s ***\n") at ../sysdeps/posix/libc_fatal.c:175 #4 0xf749c33a in malloc_printerr (action=<optimized out>, str=0xf758a9ed "free(): invalid pointer", ptr=0xb373918) at malloc.c:4996 #5 0xf749cfad in _int_free (av=0xf75d3420 <main_arena>, p=<optimized out>, have_lock=0) at malloc.c:3840 #6 0xf76799d9 in operator delete (ptr=0xb373918) at /valve/build/targets/src/gcc-4.6.1/libstdc++-v3/libsupc++/del_op.cc:47 #7 0xf7679a23 in operator delete[] (ptr=0xb373918) at /valve/build/targets/src/gcc-4.6.1/libstdc++-v3/libsupc++/del_opv.cc:32 #8 0xeb5e1b4c in PlayerManager::~PlayerManager() () from /home/tf2/tf2/tf/addons/sourcemod/bin/sourcemod.2.tf2.so #9 0xf745c543 in __cxa_finalize (d=0xeb62e200) at cxa_finalize.c:56 #10 0xeb5c3c63 in __do_global_dtors_aux () from /home/tf2/tf2/tf/addons/sourcemod/bin/sourcemod.2.tf2.so #11 0xf7731a5c in _dl_close_worker (map=map@entry=0xb312778) at dl-close.c:272 #12 0xf7732491 in _dl_close (_map=0xb312778) at dl-close.c:773 #13 0xf76c7dc2 in dlclose_doit (handle=0xb312778) at dlclose.c:35 #14 0xf772cc26 in _dl_catch_error (objname=0x90c1014, errstring=0x90c1018, mallocedp=0x90c1010, operate=0xf76c7da0 <dlclose_doit>, args=0xb312778) at dl-error.c:187 #15 0xf76c837c in _dlerror_run ( operate=operate@entry=0xf76c7da0 <dlclose_doit>, args=args@entry=0xb312778) at dlerror.c:163 #16 0xf76c7df8 in __dlclose (handle=0xb312778) at dlclose.c:46 #17 0xf3301ea4 in UnloadInterface_MMS () from /home/tf2/tf2/tf/addons/sourcemod/bin/sourcemod_mm_i486.so #18 0x0b312778 in ?? () Backtrace stopped: previous frame inner to this frame (corrupt stack?) It looks like SourceMod's player manager is trying to delete an invalid pointer.

Ive been having this happen for a good couple months now Powerlord. I dread an update because at least 75% of my servers hang on updates with these weird looking erorrs.

[MaloModo]

Originally Posted by ocwoody Disabling all lunchbox items on my servers fixed my issue. Haven't received one crash since I've done so. I run both linux and windows servers.

Unfortunately not here....I think valve needs to get to work?

[erwinquek]

Originally Posted by abrandnewday What most people experiencing crashes should do is as follows: 1. Head over to http://www.sourcemod.net/downloads.php?branch=stable and grab the latest branch for your server's OS and extract it to a folder. 2. Head over to http://metamodsource.net/ and download the "MM:S 1.10.6" download for your server's OS and extract it into the same folder. 3. Head over to http://metamodsource.net/snapshots and download the latest "1.10, stable branch" release for your server's OS. Extract that into the same folder, selecting to overwrite any already-existing files of the same name. This updates the Metamod 1.10.6 install from Step 2 to the latest version. 4. In the folder containing the SourceMod and Metamod install, navigate to "addons/sourcemod/" and delete the "configs" folder because you don't want to overwrite your admin/database/adminmenu configs. 5. If you use TF2Items: Head over to https://builds.limetech.org/?p=tf2items and download the latest version for your particular OS and extract it into the folder with the SM/MM install 6. Make sure any extensions used are up to date as well. Here's some up-to-date links for popular extensions as of the time of this post: Socket, SteamTools, SteamWorks 7. Once that's all done, simply copy over all the files onto your server's files, overwriting any existing copies when asked. Make sure to go through the stock SourceMod plugins and re-disable any that you already had disabled before the update. 8. Install Accelerator (grab the latest build from http://builds.limetech.org/?project=accelerator ) 9. On your server, navigate to addons/sourcemod/configs and open up "core.cfg". At the bottom, before the final }, add in the required "MinidumpAccount" "76561197987819599" line, changing that SteamID64 to yours (you can find it by finding yourself on SteamRep.com, it'll appear in the code block). 10. Start the server up and see if it crashes again. If it still does, head over to https://crash.limetech.org/ , login via Steam and check the logs to see if the crash was logged. If so, post the link for us to look at along with the resulting information from doing "rcon sm plugins list" and "rcon mm exts list" in the console (make sure you log into rcon on your server via "rcon_address <ip:port>" and "rcon_password <password>") Also check the latest error log in addons/sourcemod/logs and look for any errors as those may be of help. If errors are present and are from the general space of time before the crash, post them. Showing us what plugins and extensions you're using helps us. It especially helps me, because I tend to find the plugin threads here on AlliedMods and check the last page of the thread to see if people have reported any crashing issues. The other thing you can do after updating your SourceMod/Metamod install is unload plugins and move them to the disabled folder one by one until the crashes stop. Once you've determined what plugin removal made the crashes stop, that should indicate to you that it was the one causing the crashes. If anyone else reading this notices any misinformation in this post, please let me know so I can correct it. Thanks!

Thanks for the detailed sharing.

Do you see a need to disable "lunchbox items"? (for Windows)