Recent upgrade causes daily crashes

Hi.
A few days ago I upgraded a server to the most recent version of ISPConfig. Now we experience dead websites multiple times daily. It seems like Apache crashes and gets back on its feet within a few minutes. Whether this is in fact related to ISPConfig, I'm not sure. Could also be related to OS updates.

OS: Debian 12
ISPConfig: ISPConfig 3.3.1p1

Server state in ISPConfig shows all green:
State: ok (0 unknown, 0 info, 0 warning, 0 critical, 0 error)
Ok
The state of your Hard-Disk space is ok [More...]
Your virus protection is ok [More...]
Your Mail queue load is ok [More...]
Your Server load is ok [More...]
All needed services are online [More...]
Your system is up to date. [More...]
The System Log is O.K. [More...]

Restarting Apache (/etc/init.d/apache2 restart) immediately brings our websites back online, which leads me to believe this is related to Apache.

Around the time of the most recent crash (May 27 - 07:38 to 07:40), I see the following log entries (result of:
journalctl -u apache2 --since "today" --no-pager)

May 27 00:00:51 HOSTNAME systemd[1]: Reloading apache2.service - The Apache HTTP Server...
May 27 00:00:51 HOSTNAME apachectl[3995003]: [Wed May 27 00:00:51.629800 2026] [so:warn] [pid 3995003:tid 3995003] AH01574: module php_module is already loaded, skipping
May 27 00:00:51 HOSTNAME apachectl[3995003]: AH00548: NameVirtualHost has no effect and will be removed in the next release /etc/apache2/sites-enabled/000-ispconfig.vhost:7
May 27 00:00:51 HOSTNAME systemd[1]: Reloaded apache2.service - The Apache HTTP Server.
May 27 07:38:57 HOSTNAME systemd[1]: Stopping apache2.service - The Apache HTTP Server...
May 27 07:38:57 HOSTNAME apachectl[85089]: [Wed May 27 07:38:57.542411 2026] [so:warn] [pid 85089:tid 85089] AH01574: module php_module is already loaded, skipping
May 27 07:38:57 HOSTNAME apachectl[85089]: AH00548: NameVirtualHost has no effect and will be removed in the next release /etc/apache2/sites-enabled/000-ispconfig.vhost:7
May 27 07:38:59 HOSTNAME systemd[1]: apache2.service: Killing process 3995444 (PassengerAgent) with signal SIGKILL.
May 27 07:38:59 HOSTNAME systemd[1]: apache2.service: Killing process 85090 (PassengerAgent) with signal SIGKILL.
May 27 07:38:59 HOSTNAME systemd[1]: apache2.service: Killing process 3995469 (PassengerAgent) with signal SIGKILL.
May 27 07:38:59 HOSTNAME systemd[1]: apache2.service: Deactivated successfully.
May 27 07:38:59 HOSTNAME systemd[1]: Stopped apache2.service - The Apache HTTP Server.
May 27 07:38:59 HOSTNAME systemd[1]: apache2.service: Consumed 6h 47min 56.872s CPU time.
May 27 07:38:59 HOSTNAME systemd[1]: Starting apache2.service - The Apache HTTP Server...
May 27 07:38:59 HOSTNAME apachectl[85099]: [Wed May 27 07:38:59.563474 2026] [so:warn] [pid 85099:tid 85099] AH01574: module php_module is already loaded, skipping
May 27 07:38:59 HOSTNAME apachectl[85099]: AH00548: NameVirtualHost has no effect and will be removed in the next release /etc/apache2/sites-enabled/000-ispconfig.vhost:7
May 27 07:39:00 HOSTNAME systemd[1]: Started apache2.service - The Apache HTTP Server.

Could PassengerAgent be the cause ?

If I run `passenger-status`, I get the following error:

Version : 6.0.17
Date : 2026-05-27 08:19:01 +0000
Instance: GFeO58QF (Apache/2.4.67 (Debian) mod_fastcgi/mod_fastcgi-SNAP-0910052141 mod_fcgid/2.3.9 Phusion_Passenger/6.0.17)

/usr/lib/ruby/vendor_ruby/phusion_passenger/admin_tools/instance.rb:94:in `initialize': too long unix socket path (116bytes given but 108bytes max) (ArgumentError)
from /usr/lib/ruby/vendor_ruby/phusion_passenger/admin_tools/instance.rb:94:in `new'
from /usr/lib/ruby/vendor_ruby/phusion_passenger/admin_tools/instance.rb:94:in `http_request'
from /usr/sbin/passenger-status:120:in `show_status'
from /usr/sbin/passenger-status:58:in `command_show_status'
from /usr/sbin/passenger-status:366:in `start'
from /usr/sbin/passenger-status:369:in `<main>'

Any advise would be greatly appreciated - thank you

Best,
Jimmy

 

Hm, I realise that the SIGKILLs are probably related to me restarting Apache..

 

Nothing too crazy in relation to ressources:

 

Linux UNIX sockets have a hard pathname limit of 108 bytes. Passenger is trying to create or access a socket whose full filesystem path exceeds that limit.

The apache log does not show any crashes only the restart.

 

Thanks @till and @pyte. At least that rules out a few options. I have set up a monitoring script to gather more information - hopefully that will reveal the cause in the coming days. Thanks.

 

May check the journal without specifying a service name and see what happens on the server around the time it crashes - not explicitly the apache2 service but all services. In some cases you will find the cause to be something completely different.

 

@pyte Good point, I'll do that. Thanks