Analyse Apache2 coredump

Question

On my AWS server i have a lot of coredump : [Tue Nov 18 11:15:12 2014] [notice] child pid 5663 exit signal Bus error (7), possible coredump in /etc/apache2

I tried to analyse them but i got no information about it. For get it i run : sudo gdb attach pid (gdb) c (wait)

My output is :

Program received signal SIGBUS, Bus error.
0x00007f4c6a7fe1d2 in lex_scan () from /usr/lib/apache2/modules/libphp5.so
(gdb) bt
#0  0x00007f4c6a7fe1d2 in lex_scan () from /usr/lib/apache2/modules/libphp5.so
#1  0x00007f4c6a8163d0 in ?? () from /usr/lib/apache2/modules/libphp5.so
#2  0x00007f4c6a7f9795 in ?? () from /usr/lib/apache2/modules/libphp5.so
#3  0x00007f4c6a7fda7f in compile_file () from /usr/lib/apache2/modules/libphp5.so
#4  0x00007f4c6a6ba932 in ?? () from /usr/lib/apache2/modules/libphp5.so
#5  0x00007f4c6a7fdc3c in compile_filename () from /usr/lib/apache2/modules/libphp5.so
#6  0x00007f4c6a861bca in ?? () from /usr/lib/apache2/modules/libphp5.so
#7  0x00007f4c6a84f9db in execute () from /usr/lib/apache2/modules/libphp5.so
#8  0x00007f4c6a82af40 in zend_execute_scripts () from /usr/lib/apache2/modules/libphp5.so
#9  0x00007f4c6a7d75b3 in php_execute_script () from /usr/lib/apache2/modules/libphp5.so
#10 0x00007f4c6a8ba44d in ?? () from /usr/lib/apache2/modules/libphp5.so
#11 0x00007f4c6e4ff508 in ap_run_handler ()
#12 0x00007f4c6e4ff97e in ap_invoke_handler ()
#13 0x00007f4c6e50f5b0 in ap_process_request ()
#14 0x00007f4c6e50c3d8 in ?? ()
#15 0x00007f4c6e505fa8 in ap_run_process_connection ()
#16 0x00007f4c6e514210 in ?? ()
#17 0x00007f4c6e51497a in ?? ()
#18 0x00007f4c6e515527 in ap_mpm_run ()
#19 0x00007f4c6e4ea4a4 in main ()

And full backtrace :

(gdb)

bt full
#0  0x00007f4c6a7fe1d2 in lex_scan () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#1  0x00007f4c6a8163d0 in ?? () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#2  0x00007f4c6a7f9795 in ?? () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#3  0x00007f4c6a7fda7f in compile_file () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#4  0x00007f4c6a6ba932 in ?? () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#5  0x00007f4c6a7fdc3c in compile_filename () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#6  0x00007f4c6a861bca in ?? () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#7  0x00007f4c6a84f9db in execute () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#8  0x00007f4c6a82af40 in zend_execute_scripts () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#9  0x00007f4c6a7d75b3 in php_execute_script () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#10 0x00007f4c6a8ba44d in ?? () from /usr/lib/apache2/modules/libphp5.so
No symbol table info available.
#11 0x00007f4c6e4ff508 in ap_run_handler ()
No symbol table info available.
#12 0x00007f4c6e4ff97e in ap_invoke_handler ()
No symbol table info available.
#13 0x00007f4c6e50f5b0 in ap_process_request ()
No symbol table info available.
#14 0x00007f4c6e50c3d8 in ?? ()
No symbol table info available.
#15 0x00007f4c6e505fa8 in ap_run_process_connection ()
No symbol table info available.
#16 0x00007f4c6e514210 in ?? ()
No symbol table info available.
#17 0x00007f4c6e51497a in ?? ()
---Type <return> to continue, or q <return> to quit---
No symbol table info available.
#18 0x00007f4c6e515527 in ap_mpm_run ()
No symbol table info available.
#19 0x00007f4c6e4ea4a4 in main ()
No symbol table info available.

How can i get detailed information about this trace ?

Answer 1

I tried to analyse them but i got no information about it.

On the contrary, you did get information about it. As Mark Plotnick said, to get more detailed info, you'll need to install debuginfo packages for your distribution.

The crash looks awfully similar to this one.

Answer 2

One reason found on a Ubuntu 16 / PHP 7 / apache 2.4 system was a certificate mismatch:

After an apache restart, the log file XXX.log was overflooded with messages like

[core:notice] child pid nnnnn exit signal bus error 7

The site went effectively unusable due to the many log messages.

Searching the setup it showed that no enabled site was setup to use XXX.log, but in the apache2/sites-available directory a file named "default-tls.conf" (belongs to package "libapache2-mod-gnutls") defined XXX.log as it's error log file.

Though that conf-file was not enabled, not linked into apache2/sites-enabled, the XXX.log was flooded.

The system uses for long time a letsencrypt cert, all enabled sites were setup to use that. But in default-tls.conf cert and key files were setup to point to the (present) default "snakeoil" files. The "snakeoil" cert and key files were replaced by links to the letsencrypt cert and key files.

After an apache restart the error messages were not more thrown anymore and the websites were accessable again.

What helped to find the reason was the use of a different from standard error log file setup in the vhost conf files.