Issue Segmentation fault (11) Centos 7

[Kabamaru]

Hello everyone,

I recently migrated my plesk install CentOS 6 centos 7 without problems.

Suddenly a week ago websites have stopped working and the file "/var/log/httpd/error_log" have appeared many lines like this:

Fri Jul 01 20:00:05.466270 2016] [core:notice] [pid 12871:tid 140517709097024] AH00051: child pid 933 exit signal Segmentation fault (11)

I have enabled the coredump and this is the result:

gdb /usr/sbin/httpd core-\!usr\!sbin\!httpd-11-48-48-15895-1467408614
GNU gdb (GDB) Red Hat Enterprise Linux 7.6.1-80.el7
Copyright (C) 2013 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 "x86_64-redhat-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/sbin/httpd...Reading symbols from /usr/lib/debug/usr/sbin/httpd.debug...done.
done.

warning: core file may not match specified executable file.
[New LWP 15916]
[New LWP 15919]
[New LWP 15917]
[New LWP 15911]
[New LWP 15918]
[New LWP 15910]
[New LWP 15895]
[New LWP 15920]
[New LWP 15908]
[New LWP 15897]
[New LWP 15922]
[New LWP 15915]
[New LWP 15898]
[New LWP 15904]
[New LWP 15921]
[New LWP 15900]
[New LWP 15899]
[New LWP 15906]
[New LWP 15914]
[New LWP 15901]
[New LWP 15903]
[New LWP 15913]
[New LWP 15902]
[New LWP 15912]
[New LWP 15907]
[New LWP 15905]
[New LWP 15909]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
Core was generated by `/usr/sbin/httpd -DFOREGROUND'.
Program terminated with signal 11, Segmentation fault.
#0 __GI___pthread_mutex_lock (mutex=mutex@entry=0x41) at pthread_mutex_lock.c:50
50 unsigned int type = PTHREAD_MUTEX_TYPE (mutex);
(gdb) bt
#0 __GI___pthread_mutex_lock (mutex=mutex@entry=0x41) at pthread_mutex_lock.c:50
#1 0x00007fccc4cff58a in modperl_tipool_putback_base (tipool=0x41, listp=listp@entry=0x0, data=data@entry=0x7fccd7941550, num_requests=2001) at modperl_tipool.c:297
#2 0x00007fccc4cff89c in modperl_tipool_putback_data (tipool=<optimized out>, data=data@entry=0x7fccd7941550, num_requests=<optimized out>) at modperl_tipool.c:393
#3 0x00007fccc4cfecd8 in modperl_interp_unselect (data=0x7fccd7941550) at modperl_interp.c:299
#4 0x00007fccd2f4c1ae in run_cleanups (cref=<optimized out>) at memory/unix/apr_pools.c:2352
#5 apr_pool_destroy (pool=0x7fcca8028d78) at memory/unix/apr_pools.c:814
#6 0x00007fccd428db06 in remove_empty_buckets (bb=bb@entry=0x7fccb8005988) at core_filters.c:716
#7 0x00007fccd428dea8 in send_brigade_nonblocking (s=0x7fccb8004f00, bb=bb@entry=0x7fccb8005988, bytes_written=bytes_written@entry=0x7fccb8005940, c=c@entry=0x7fccb8005180) at core_filters.c:705
#8 0x00007fccd428ed5a in ap_core_output_filter (f=0x7fccb80057e8, new_bb=0x0) at core_filters.c:469
#9 0x00007fccc986172f in process_socket (my_thread_num=19, my_child_num=<optimized out>, cs=0x7fccb8005108, sock=<optimized out>, p=<optimized out>, thd=<optimized out>) at event.c:983
#10 worker_thread (thd=<optimized out>, dummy=<optimized out>) at event.c:1812
#11 0x00007fccd2d1edc5 in start_thread (arg=0x7fccb07f0700) at pthread_create.c:308
#12 0x00007fccd2847ced in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:113
(gdb)

I have no idea why, because it had not failed before.

Thanks in advance for any help.

 

[UFHH01]

Hi Kabamaru,

consider to use MPM prefork instead of MPM worker and report back, if that already solves your issue.

 

[Kabamaru]

I do not know if it will be the same, but only got as options "Event" and "Prefork" but I have not the "Worker" option. By default I have selected the "Event" option, so I will change to "Prefork" and will check if it solves.

A question about the change: Are not supposed to be better "Event" that "Prefork"? Source: http://serverfault.com/questions/383526/how-do-i-select-which-apache-mpm-to-use (not saying they're right), because he had thought make your suggestion before, but it indicating on the web I did not realize it.

I do not remember how it was previously set to Centos 6, so I can not compare.

Thank you very much.

 

[UFHH01]

Hi Kabamaru,

"event" and "worker" are pretty similar, especially when it comes to threading, but I noticed as well in your coredump the mod_perl usage. You have to recompile perl with the "-Dusethreads" option, so perl is able to properly support "event" and "worker" with mod_perl - usage. This is no standart and most people just install the vendor's version.
"Prefork" is the most stable option, with the highest compatibility and is recommended when you use FastCGI or PHP-FPM.

For further investigations on your issue, pls. consider to post your apache version, your php-handler and the compiling options. It would not be bad to post as well the used apache-modules and when you add the corresponding log - file entries for your issues from your domains.