zend_mm_heap corrupted and segfaults

[Richard G]

I'm encountering loads of this in the apache error log of one server:

[Wed Sep 30 20:56:07.128271 2020] [core:notice] [pid 26048] AH00052: child pid 27277 exit signal Segmentation fault (11)
Wed Sep 30 20:56:07 2020 (26912): Error Cannot kill process 26920: Operation not permitted!
[Wed Sep 30 20:56:08.134283 2020] [core:notice] [pid 26048] AH00052: child pid 26912 exit signal Segmentation fault (11)
Wed Sep 30 20:56:13 2020 (28123): Error Cannot kill process 26920: Operation not permitted!
Wed Sep 30 20:56:15 2020 (28124): Error Cannot kill process 26920: Operation not permitted!
Wed Sep 30 20:56:20 2020 (28125): Error Cannot kill process 26920: Operation not permitted!
Wed Sep 30 20:56:21 2020 (28127): Error Cannot kill process 26920: Operation not permitted!
Wed Sep 30 20:56:24 2020 (28074): Error Cannot kill process 26920: Operation not permitted!
Wed Sep 30 20:56:47 2020 (28080): Error Cannot kill process 26920: Operation not permitted!
Wed Sep 30 20:56:48 2020 (28105): Error Cannot kill process 26920: Operation not permitted!
Wed Sep 30 20:56:55 2020 (28111): Error Cannot kill process 26920: Operation not permitted!
Wed Sep 30 20:56:55 2020 (28123): Error Cannot kill process 26920: Operation not permitted!
zend_mm_heap corrupted
Wed Sep 30 20:56:58 2020 (28125): Error Cannot kill process 26920: Operation not permitted!
zend_mm_heap corrupted

I did install opcache on all servers, but only this servers seems to have an issue. I did not enable opcache in the php.ini.
Furthermore I would like to test by removing opcache, but things like:
./build remove_items
or
./build remove_opcache
are not working.

On this server we experienced odd outages. Blank screens on websites, for example Wordpress gives a blank screen, WP-admin is accessible. After restarting httpd and mysql everything works fine again.
Very strange problem.

 

[factor]

What does "zend_mm_heap corrupted" mean

All of the sudden I've been having problems with my application that I've never had before. I decided to check the Apache's error log, and I found an error message saying "zend_mm_heap corrupted"....

stackoverflow.com

 

[Richard G]

So where can I set this:
opcache.fast_shutdown=0
Because I don't see a opcache.fast_shutdown=1 in php.ini.

Do I need to add this manually to the php.ini?

 

[factor]

PHP: Runtime Configuration - Manual

PHP is a popular general-purpose scripting language that powers everything from your blog to the most popular websites in the world.

www.php.net

 

[factor]

try here
cd /usr/local/directadmin/custombuild/configure/opcache

 

[factor]

yep
cat opcache.ini
opcache.memory_consumption=128
opcache.interned_strings_buffer=8
opcache.max_accelerated_files=4000
opcache.revalidate_freq=0
opcache.enable_cli=1
opcache.validate_permission=1
opcache.use_cwd=1
opcache.revalidate_path=1
opcache.enable_file_override=1

 

[factor]

What does "zend_mm_heap corrupted" mean

All of the sudden I've been having problems with my application that I've never had before. I decided to check the Apache's error log, and I found an error message saying "zend_mm_heap corrupted"....

stackoverflow.com

the post seems to point to older code with memory leaks. looks like there are several things to try. if you want to.

 

[factor]

On this server we experienced odd outages.

is this all after using opcache or in general?

 

[factor]

20421 – child pid $PID exit signal Segmentation fault (11)

bz.apache.org

for apache are you using mpm event or something else?

 

[Richard G]

is this all after using opcache or in general?

This was after installing opcache as far as I remember.
On this server I uninstalled opcache and all seems quiet now again, at least with the errors. I can reinstall it and then see if it happens again with changed opcache settings, but I rather not.
The other servers are running Opcache too, so I will put that line in there to see what it does.

for apache are you using mpm event or something else?

I've seen mpm prefork in the log. Server is running on mod_php with mod_ruid2.

 

[Richard G]

Added:
opcache.fast_shutdown=0
tot the opcache.ini on the other server, there was no line like that present yet. Restarted apache, hope it will keep things fine.

 

[factor]

This was after installing opcache as far as I remember.
On this server I uninstalled opcache and all seems quiet now again, at least with the errors. I can reinstall it and then see if it happens again with changed opcache settings, but I rather not.
The other servers are running Opcache too, so I will put that line in there to see what it does.


I've seen mpm prefork in the log. Server is running on mod_php with mod_ruid2.

httpd -V | grep 'MPM'

• apache_mpm - this option is used to choose the MPM (Multi-Processing Module) of Apache. MPM is chosen during "./build apache". Possible values: prefork, event, auto (default: auto). Auto option selects "Event" MPM if mod_php is not chosen as php1_mode, otherwise "Prefork" MPM is set.

 

[Richard G]

httpd -V | grep 'MPM'

Server MPM:     prefork

I have mod_php choosen as php1_mode (we only run 1 mode due to mod_ruid2) so only prefork is used.