| SYSTEM INFORMATION | |
|---|---|
| OS type and version | Ubuntu 20.04 |
| Webmin version | 2.001 |
| Virtualmin version | 7.2-1-Pro |
| Related packages | Systemd 245 |
Hi all,
Does anyone know why systemd only kills the user slices of Virtialmin virtual server admins?
First I got this error message for almost all my Vmin admin users (corresponding id’s):
user@1078.service loaded failed failed User Manager for UID 1078
So from there I restarted a service
# systemctl status user@1078
● user@1078.service - User Manager for UID 1078
Loaded: loaded (/lib/systemd/system/user@.service; static; vendor p>
Drop-In: /usr/lib/systemd/system/user@.service.d
└─timeout.conf
Active: active (running) since Wed 2022-10-12 13:57:46 CST; 3s ago
Docs: man:user@.service(5)
Main PID: 3323694 (systemd)
Status: "Startup finished in 241ms."
Tasks: 6
Memory: 6.6M
CGroup: /user.slice/user-1078.slice/user@1078.service
├─dbus.service
│ └─3323708 /usr/bin/dbus-daemon --session --address=system>
├─init.scope
│ ├─3323694 /lib/systemd/systemd --user
│ └─3323700 (sd-pam)
└─pulseaudio.service
└─3323705 /usr/bin/pulseaudio --daemonize=no --log-targe
And I ran # killsnoop-bpfcc -p 3323694 in order to determine is killing the prcoess:
TIME PID COMM SIG TPID RESULT
14:01:22 3323694 systemd 15 3323708 0
14:01:22 3323694 systemd 18 3323708 0
In essence there seems to be something about Vmin admin users running php apps (wordpress, roundcube, & phpmyadmin) that Systemd doesn’t like… and therefore kills the process.
The process functioning properly it appears as follows:
● user@1078.service - User Manager for UID 1078
Loaded: loaded (/lib/systemd/system/user@.service; static; vendor preset: enabled)
Drop-In: /usr/lib/systemd/system/user@.service.d
└─timeout.conf
Active: active (running) since Wed 2022-10-12 14:37:58 CST; 11s ago
Docs: man:user@.service(5)
Main PID: 3337802 (systemd)
Status: "Startup finished in 244ms."
Tasks: 6
Memory: 6.7M
CGroup: /user.slice/user-1078.slice/user@1078.service
├─dbus.service
│ └─3337816 /usr/bin/dbus-daemon --session --address=systemd: --nofork --nopidfile --systemd-activation --syslog-only
├─init.scope
│ ├─3337802 /lib/systemd/systemd --user
│ └─3337808 (sd-pam)
└─pulseaudio.service
└─3337813 /usr/bin/pulseaudio --daemonize=no --log-target=journal
Oct 12 14:37:58 admin.example.com systemd[3337802]: Listening on D-Bus User Message Bus Socket.
Oct 12 14:37:58 admin.example.com systemd[3337802]: Reached target Sockets.
Oct 12 14:37:58 admin.example.com systemd[3337802]: Reached target Basic System.
Oct 12 14:37:58 admin.example.com systemd[1]: Started User Manager for UID 1078.
Oct 12 14:37:58 admin.example.com systemd[3337802]: Starting Sound Service...
Oct 12 14:37:58 admin.example.com systemd[3337802]: Started D-Bus User Message Bus.
Oct 12 14:37:58 admin.example.com dbus-daemon[3337816]: [session uid=1078 pid=3337816] AppArmor D-Bus mediation is enabled
Oct 12 14:37:58 admin.example.com systemd[3337802]: Started Sound Service.
Oct 12 14:37:58 admin.example.com systemd[3337802]: Reached target Main User Target.
Oct 12 14:37:58 admin.example.com systemd[3337802]: Startup finished in 244ms.
NOTE:
Most of my Vmin admin users have existed for months on the server without this issue happening. Only recently, have I begun adding php applications, which seem to have to pissing off Systemd.
Anyone have an ideas on how to Systemd and Vmin playing together nicely?
