Issue Error after upgrading to php 7.2.6 from php 7.2.5

[Nicola]

Hi,
after the upgrade I get a lot of this errors in Apache error log
"(104)Connection reset by peer: AH01075: Error dispatching request to :, referer: htps://....".

At the same time in "plesk-php72-fpm/error.log" I get the error
"WARNING: [pool www.xxxx.xx] child 156685 exited on signal 11 (SIGSEGV) after 143.747383 seconds from start"

I tried to disable/change value of params "MaxConnectionsPerChild"/"pm.max_requests" but this not resolve.

Thanks

 

[Bitpalast]

Maybe you need to restart the service
# service plesk-php72-fpm restart

 

[Nicola]

Already tried several times, but it does not work. I noticed that the error also occurs with php 7.1

 

[Bitpalast]

Can you access your website normally or does PHP 7.x not work at all?

 

[Nicola]

I can access normally, errors occur sometimes every minute and cause 503 errors in users' browsers. It seems as if the processes are terminated too quickly.

 

[Bitpalast]

My suggestion:

Please try these values:


The number of max_children is an example and should be higher if you see a max children warning in the /var/log/plesk-php7x-fpm/error.log files.

After the change, restart the PHP-FPM service of the corresponding version, for example
# service plesk-php72-fpm restart

If the issue persists afterwards, please backup /etc/httpd/conf.d/fcgid.conf apply changes to /etc/httpd/conf.d/fcgid.conf:

FcgidIdleTimeout 20
FcgidProcessLifeTime 20
FcgidMaxProcesses 800
FcgidMaxProcessesPerClass 30
FcgidIOTimeout 20

For IdleTimeout, ProcessLifeTime and IOTimeout the values should be decreased from the default., For MaxProcesses and MaxProcessesPerClass the values should be increased, but not necessarily to exactly the values given in this example. It depends on how much RAM your system has. Normally, the given values should be o.k. for a system that uses 128 GB or more.

After such changes, please restart Apache
# apachectl -t
# service httpd restart

If the issue persists afterwards, it is very likely not caused by PHP-FPM, but by a script of the website that does not respond to certain requests. This can be either caused by the script or by a ModSecurity block or by missing local IPv4 or 127.0.0.1 entries in Fail2Ban.

 

[Nicola]

If I change php-fpm to php-FastCGI the problem does not occur. The problem remains with php-fpm

 

[Bitpalast]

Have you tried the recommended changes?

 

[Nicola]

Yes, I tried to change the file but I do not have any improvments.

 

[Bitpalast]

Do you see any error messages in the corresponding PHP-FPM log (for example /var/log/plesk-php71-fpm/error.log or /var/log/plesk-php72-fpm/error.log) that do not refer to program source code errors but FPM errors? It is very unlikely that with the settings from above there remain issues with the service, but there should be clear error messages in the log files if there remain issues with FPM settings.

 

[Nicola]

Thi is an example of errors that I see in /var/log/plesk-php72-fpm/error.log

[30-May-2018 10:40:40] WARNING: [pool example.domain.tld] child 120903 exited on signal 11 (SIGSEGV) after 406.135331 seconds from start
[30-May-2018 10:40:40] NOTICE: [pool example.domain.tld] child 128705 started
[30-May-2018 10:40:45] ERROR: unable to read what child say: Bad file descriptor (9)
[30-May-2018 10:40:45] ERROR: unable to read what child say: Bad file descriptor (9)
[30-May-2018 10:40:59] WARNING: [pool example.domain.tld] child 126393 exited on signal 11 (SIGSEGV) after 131.591423 seconds from start
[30-May-2018 10:40:59] NOTICE: [pool example.domain.tld] child 128968 started
[30-May-2018 10:41:03] WARNING: [pool example.domain.tld] child 121882 exited on signal 11 (SIGSEGV) after 369.867722 seconds from start
[30-May-2018 10:41:03] NOTICE: [pool example.domain.tld] child 129287 started
[30-May-2018 10:41:29] WARNING: [pool example.domain.tld] child 110084 exited on signal 11 (SIGSEGV) after 999.956164 seconds from start
[30-May-2018 10:41:29] NOTICE: [pool example.domain.tld] child 129650 started
[30-May-2018 10:41:43] WARNING: [pool example.domain.tld] child 102250 exited on signal 11 (SIGSEGV) after 1369.911259 seconds from start
[30-May-2018 10:41:43] NOTICE: [pool example.domain.tld] child 129872 started
[30-May-2018 10:41:54] WARNING: [pool example.domain.tld] child 124110 exited on signal 11 (SIGSEGV) after 314.605024 seconds from start
[30-May-2018 10:41:54] NOTICE: [pool example.domain.tld] child 129936 started
[30-May-2018 10:42:56] WARNING: [pool example.domain.tld] child 112246 exited on signal 11 (SIGSEGV) after 967.565714 seconds from start
[30-May-2018 10:42:56] NOTICE: [pool example.domain.tld] child 131078 started
[30-May-2018 10:43:07] WARNING: [pool example.domain.tld] child 119637 exited on signal 11 (SIGSEGV) after 620.962504 seconds from start
[30-May-2018 10:43:07] NOTICE: [pool example.domain.tld] child 131585 started
[30-May-2018 10:43:33] WARNING: [pool example.domain.tld] child 122248 exited on signal 11 (SIGSEGV) after 510.539898 seconds from start
[30-May-2018 10:43:33] NOTICE: [pool example.domain.tld] child 131841 started
[30-May-2018 10:43:42] WARNING: [pool example.domain.tld] child 86164 exited on signal 11 (SIGSEGV) after 2182.220712 seconds from start
[30-May-2018 10:43:42] NOTICE: [pool example.domain.tld] child 131878 started
[30-May-2018 10:43:49] WARNING: [pool example.domain.tld] child 128968 exited on signal 11 (SIGSEGV) after 170.021542 seconds from start
[30-May-2018 10:43:49] NOTICE: [pool example.domain.tld] child 131963 started
[30-May-2018 10:44:11] WARNING: [pool example.domain.tld] child 118323 exited on signal 11 (SIGSEGV) after 747.359966 seconds from start
[30-May-2018 10:44:11] NOTICE: [pool example.domain.tld] child 132564 started
[30-May-2018 10:44:12] WARNING: [pool example.domain.tld] child 120419 exited on signal 11 (SIGSEGV) after 652.240825 seconds from start
[30-May-2018 10:44:12] NOTICE: [pool example.domain.tld] child 132566 started
[30-May-2018 10:44:15] ERROR: unable to read what child say: Bad file descriptor (9)
[30-May-2018 10:44:15] ERROR: unable to read what child say: Bad file descriptor (9)
[30-May-2018 10:44:15] WARNING: [pool example.domain.tld] child 113406 exited on signal 11 (SIGSEGV) after 982.968256 seconds from start
[30-May-2018 10:44:15] NOTICE: [pool example.domain.tld] child 132734 started
[30-May-2018 10:44:40] WARNING: [pool example.domain.tld] child 131878 exited on signal 11 (SIGSEGV) after 58.480803 seconds from start
[30-May-2018 10:44:40] NOTICE: [pool example.domain.tld] child 133178 started

 

[Bitpalast]

Thank you for sharing. I have not yet seen a similar situation on another server, so I cannot say exactly why it occurs. For me all similar symptoms could always be solved by changes to the PHP-PFM configuration.

Some people say, that is is caused by a "wrong" order of PHP extensions in PHP, and more frequently it is mentioned that "New Relic" is causing this symptom (don't bother if you have not installed it). Some other say, it is opcache that is causing it. Unfortunately this is impossible to analyze further by remote advice. I cannot give further advice on this.

 

[harpreetsb]

Having exactly same issue with php7.2
I decided to reboot the server and nothing worked after it. everything on 503 since morning.

[21-Oct-2019 12:47:21] WARNING: [pool test.medigence.com] child 12367 exited on signal 11 (SIGSEGV - core dumped) after 0.204068 seconds from start
[21-Oct-2019 12:47:21] NOTICE: [pool test.medigence.com] child 12371 started
[21-Oct-2019 12:51:24] WARNING: [pool test.medigence.com] child 13027 exited on signal 11 (SIGSEGV - core dumped) after 0.194998 seconds from start
[21-Oct-2019 12:51:24] NOTICE: [pool test.medigence.com] child 13029 started
[21-Oct-2019 12:52:18] WARNING: [pool test.medigence.com] child 13252 exited on signal 11 (SIGSEGV - core dumped) after 0.186359 seconds from start
[21-Oct-2019 12:52:18] NOTICE: [pool test.medigence.com] child 13254 started
[21-Oct-2019 12:52:21] WARNING: [pool test.medigence.com] child 13254 exited on signal 11 (SIGSEGV - core dumped) after 2.124904 seconds from start
[21-Oct-2019 12:52:21] NOTICE: [pool test.medigence.com] child 13256 started
[21-Oct-2019 12:52:22] WARNING: [pool test.medigence.com] child 13256 exited on signal 11 (SIGSEGV - core dumped) after 1.286462 seconds from start
[21-Oct-2019 12:52:22] NOTICE: [pool test.medigence.com] child 13261 started

 

[mow]

Having exactly same issue with php7.2
I decided to reboot the server and nothing worked after it. everything on 503 since morning.

[21-Oct-2019 12:47:21] WARNING: [pool test.medigence.com] child 12367 exited on signal 11 (SIGSEGV - core dumped) after 0.204068 seconds from start

SIGSEGV is a sign of broken software, i.e. out-of-bounds read/write/execute which can be caused by buffer overflow or memory corruption, or a broken/mismatched build.
Could you find the core dump file and have a look with gdb for what went wrong?

 

[harpreetsb]

Well solved it by upgrading to php7.3