I recently updated my (busy) site to a new chat room and I think it's taxing the server, but I'm not sure why or how. I'm receiving an email like this from my server about once a day, which is concerning. I'm running nginx as the web server that calls php-fpm.
Subject line: [abrt] php71w-fpm: php-fpm killed by SIGSEGV
reason: php-fpm killed by SIGSEGV
cmdline: 'php-fpm: pool www' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' ''
executable: /usr/sbin/php-fpm
package: php71w-fpm-7.1.32-1.w7
component: php71w
pid: 21409
pwd: /usr/share/nginx/www.example.com/chatroom
hostname: main.example.com
count: 63
abrt_version: 2.1.11
analyzer: CCpp
architecture: x86_64
event_log:
global_pid: 21409
kernel: 3.10.0-957.21.3.el7.x86_64
last_occurrence: 1572875501
os_release: CentOS Linux release 7.6.1810 (Core)
pkg_arch: x86_64
pkg_epoch: 0
pkg_fingerprint: E87F D236 62E7 4CA5
pkg_name: php71w-fpm
pkg_release: 1.w7
pkg_vendor: Webtatic
pkg_version: 7.1.32
runlevel: N 3
time: Mon 14 Oct 2019 12:09:23 PM GMT
type: CCpp
uid: 995
username: nginx
uuid: b008d2dabbed483b51d5bc88d507b2d0288f898d
core_backtrace:
:{ "signal": 11
:, "executable": "/usr/sbin/php-fpm"
:, "stacktrace":
: [ { "crash_thread": true
: , "frames":
: [ { "address": 94687732525503
: , "build_id": "7c6a08e78d7575f9b7120f9bba3686286da0c5c6"
: , "build_id_offset": 3279295
: , "function_name": "ZEND_INCLUDE_OR_EVAL_SPEC_CV_HANDLER"
: , "file_name": "/usr/sbin/php-fpm"
: }
(and then about a million of those "frames" that all look similar)
Is this a case where php-fpm is running out of memory (or something)? How can I fix this? Is there a blog post or something that can help me "tune" my PHP to solve? Thank you...
0 Answers