on a Debian 10 server running PHP-FPM 7.3.19 installed from official Debian repos, I have some apparently random segmentation faults.
These happens on other 2 Debian 10 servers running same php version, but does not happen on:
- Ubuntu 18.04, PHP-FPM 7.2.24
- Debian 9, Php-FPM 5.6.40
My skills here are very limmited, I was able to strace
the process killed by SIGSEGV
, but I can't see a smocking gun.
I was also able to obtain a core dump, and analyze it with gdb
bt
, but I don't know how to manage this info to find the cause of the segfault.
Can somebody more skilled help me?
These is the bt
of the core dump:
#0 0x000055efb6a3ae0d in _emalloc ()
#1 0x000055efb6ada749 in ?? ()
#2 0x000055efb6adf22d in execute_ex ()
#3 0x000055efb6a518de in zend_call_function ()
#4 0x000055efb695897c in ?? ()
#5 0x000055efb6a51798 in zend_call_function ()
#6 0x000055efb6a51c90 in zend_lookup_class_ex ()
#7 0x000055efb6a52678 in zend_fetch_class_by_name ()
#8 0x000055efb6aabf16 in ?? ()
#9 0x000055efb6ae073e in execute_ex ()
#10 0x000055efb6a518de in zend_call_function ()
#11 0x000055efb695897c in ?? ()
#12 0x000055efb6a51798 in zend_call_function ()
#13 0x000055efb6a51c90 in zend_lookup_class_ex ()
#14 0x000055efb6a52678 in zend_fetch_class_by_name ()
#15 0x000055efb6ad8407 in ?? ()
#16 0x000055efb6ae0776 in execute_ex ()
#17 0x000055efb6ae6dd7 in zend_execute ()
#18 0x000055efb6a5f863 in zend_execute_scripts ()
#19 0x000055efb69fff88 in php_execute_script ()
#20 0x000055efb68c7360 in ?? ()
#21 0x00007f1a094ab09b in __libc_start_main (main=0x55efb68c65d0, argc=4, argv=0x7ffcf9e0ad58, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7ffcf9e0ad48) at ../csu/libc-start.c:308
#22 0x000055efb68c820a in _start ()