• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Issue Apache crashing every night

Ivo

Basic Pleskian
Hello

We are having a bit of issue with one of our plesk 12.5 debian 8 servers.

Apache is crashing every couple of days. Today we got this error:

[Fri Feb 03 02:00:15.076267 2017] [mpm_prefork:emerg] [pid 26879] (35)Resource deadlock avoided: AH00144: couldn't grab the accept mutex
[Fri Feb 03 02:00:16.308386 2017] [core:alert] [pid 26875] AH00050: Child 26879 returned a Fatal error... Apache is exiting!

Then I had to manually kill every apache process that was still living and start apache manually to get websites working.
Regular restart didn't work.

I did read a bit and saw that plesk 12.5 with debian has some sort of issues so I tried solutions from
https://support.plesk.com/hc/en-us/articles/213946305 but so far no luck.

Can anyone help me a bit with this issue?
 
Are you using mod_ruid2?
You could try to add one directive into your custom Apache configuration of the affected domain:

For Apache version prior 2.4 either this:
Code:
AcceptMutex flock
or this:
Code:
AcceptMutex posixsem

For Apache version 2.4 or younger either this:
Code:
Mutex flock
or this:
Code:
Mutex posixsem
 
Hello

Every domain is affected when this issues happens. I think it happens when he tries to logrotate.

How would we proceed dealing with issue globally?

Atm we have these settings:
Apache restart interval is 60 sec
piped logs are not activated
Mpm mode prefork

Can't see anywhere mod_ruid2 in modules.

Logrotate conf is set to restart instead of reload cause that was crashing sites as well.
 
Logrotate can cause Apache crashs. In other cases, a second or third restart attempt while the first or first two were not yet completed has caused Apache to crash. These additional Apache restarts could be defined in postrotate directives in service configuration files (the conf files in /etc/logrotate.d). One could be modsecurity.

However, in your case, the error message you cited is NOT indicating a crash for the reason of logrotate, but file locking failures. Maybe you would like to find out more details. Change the LogLevel directive in /etc/httpd/conf/httpd.conf from "warn" to "debug", restart Apache, then let the crash happen, then change it back to "warn" and analyze the logs. The logs should then contain more details on the crash situation.
 
Back
Top