Test System now getting Segmentation Faults

[aegersz]

hi, i have a dev. system and i notice that when i import the live database there, i see 100's of members reported as active (which is incorrect) and now i am getting quite a few [notice] child pid 30178 exit signal Segmentation fault (11)

because i rysnc from the live system and use phpmyadmin for the database copies, why do you think i am getting segmentation errors suddenly ?

i seem to have enough free memory, btw ...

any thoughts ? i can only use my dev. SMF system in Firefox and it no longer works in Chrome - another weird oddity.

[Shambles]

Have a check here

.. specifically at the question "Have you tried to increase output_buffering in your php.ini?"

[aegersz]

yes, i did try that prior to posting this.

if i take a fresh version of SMF then it's ok but for some reason, my rsync'ed from live leads me to this error.

how do you guys handle your test systems ? do you use live data and also get the sort of problems i have ?

[Shambles]

I use MySQLdumper to execute every 2 hours on my live host, sending the resulting DB copy to my test server.

On that test server, MySQLdumper uploads the data copy to my test DB.

[aegersz]

thanks, yeah, i take hourly MySQLdumper dumps also.

i will leave this open for a while so maybe someone else has any thoughts.

i don't know why i am stressing - it's only a dev system BUT i hate anomalies like this and want to know why this started to happen.

[Shambles]

You do know MySQLdumper can ftp its DB copy to your server...?

[aegersz]

no i didn't but i only rarely have my test system up so i'm coping with the manual route.

i reinstalled SMF 2.0.10 on my test system and that solved the problem.

i'd still like to know why it happens ? i think there is a dodgy mod'ed Source file (I now have 56 mods so i guess it's likely ...

am i right to assume so ?

LATER ON ... Yes, I have isolated the problem to SMF Source, but how can i find out which program is the culprit and is it really a problem ? My live system is working fine ...

[margarett]

Did you ran repair_settings after importing the DB?
It's very odd that you get a segfault, really... Wrong settings wouldn't cause that but it's worth to try...

[aegersz]

i didn't but could it be hardware related ? my dev machine is a really old laptop

[aegersz]

gdb apache2 -core /tmp/apache2-gdb-dump/core.6281
GNU gdb (GDB) Red Hat Enterprise Linux (7.2-83.el6)
Copyright (C) 2010 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "i686-redhat-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
apache2: No such file or directory.
Missing separate debuginfo for the main executable file
Try: yum --enablerepo='*-debug*' install /usr/lib/debug/.build-id/1a/fc88da2283c6a5812c2fa4539cdee2b0aed105
[New Thread 6281]
Core was generated by `/usr/sbin/httpd'.
Program terminated with signal 11, Segmentation fault.
#0  0xb765a179 in ?? ()
(gdb)

... i tried the yum --enablerepo etc. but it failed.

any more thoughts ?

[Illori]

if you really want to know why,

Missing separate debuginfo for the main executable file
Try: yum --enablerepo='*-debug*' install /usr/lib/debug/.build-id/1a/fc88da2283c6a5812c2fa4539cdee2b0aed105

then wait for the next core dump

[margarett]

I can't think of a way for SMF to crash Apache...

Do you use Pretty URLs? Maybe incorrect settings could cause a redirect loop? Dunno, just guessing...

[aegersz]

there is a loop there somewhere because the noise i get when i get a message from SA Chat (mod) keeps looping.

i just can't get any meaningful debug data.

i wish i could find out WHAT php file is causing this - i'm struggling a bit despite reviewing many core dumps but from IBM OS'es.

[Illori]

your issue is the software you are using is not build with debug symbols so you will get nothing useful out of them unless you rebuild with debug enabled.

[aegersz]

yes, i understand that but can't figure out how to do so.

it's probably easier if i DO NOT rsync Source libs from Prod to Dev. and just apply the mods as needed.

is that how you do it ? is the Source meant to be copied from one system to another or do you advise Mods to do this ?

[Illori]

i mean that your apache etc has to be built with debug in it, this is an software level build. if you need help doing that you should find a forum for your version of your OS and post their for help.

this has NOTHING to do with how you sync your SMF installs between the 2 machines.

[aegersz]

this has NOTHING to do with how you sync your SMF installs between the 2 machines.

i understand but some code that runs in Production doesn't in Test.

I have one last question for now ... do i have to enable debug in PHP or APACHE ?

[Illori]

the segfault seems to be in apache, so i would start there.

[aegersz]

i think i worked it out:

i was  also getting this:
PHP Fatal error:  Class 'Facebook' not found in /var/www/html/Sources/Facebook/FacebookHooks.php on line 419

PHP Warning:  include(): Failed opening '/var/www/html/facebookauth/facebook.php' ...

somehow, i managed to lose the facebookauth directory and once i copied it, it seems ok now ...

sorry for causing all this grief ! i'm not sure why the rsync didn't pick it up but i did remove the mod and reinstall it at some point.

i think i have been wasting your time.

[margarett]

No worries But again, a PHP fatal error should not crash Apache...