Hot answers tagged

3

The targets depend on OS, which you haven't listed here. Probably default.target. You can view targets and what uses by looking at the output of systemctl list-dependencies Just remove "Restart" entirely The initial service should be Type=oneshot. It will be up to the service to properly exit based on whether or not it wrote the file you require. You may ...


2

Usually that is ntp (the daemon named ntpd) for Network Time Protcol. For example: Network Time Protocol daemon (Arch) RHEL7: How to set up the NTP service. In Fedora, you may be looking for timedatex.service which is related to this package: timedatex is a D-Bus service that implements the org.freedesktop.timedate1 interface. It can be used to read ...


2

Try this to make your current machine ID permanent: # mount --bind / /mnt # cp /etc/machine-id /mnt/etc/machine-id # reboot This should get rid of the tmpfs mount over /etc/machine-id.


2

Can you try this as your service file. [Unit] After=wpa_supplicant.service dbus.service NetworkManager.service Requires=wpa_supplicant.service dbus.service NetworkManager.service [Service] Type=oneshot RemainAfterExit=true ExecStart=/bin/true ExecStop=MyShutdownExecutable [Install] wantedBy=multi-user.target I managed to get my script running with this ...


1

ok. SELinux issue. When creating some new stuff related to existing service, you need to make sure that the service will have appropriate access to your files. The log from your file proposes that it does not have (unless running as unconfined_t). executing action "/etc/acpi/actions/lid.sh "button/lid LID open"" action exited with status 126 The execution ...


1

Alright, being able to start the service manually sounds good. You also have to enable the service with systemctl enable radio? From the systemctl status radio it looks like the service is not enabled and thus doesn't start.


1

The unit runs on system boot because it is enabled. You want it to be started by the timer unit only.


1

I had the same issue. The debian maintainers added a patch to mdadm which causes the raid assembly to start before the devices are up. I have not yet completely found out why as it's supposed to fix broken RAIDs as root file system. But you can fix it for now by downgrading the mdadm package. Get the older version from here: ...


1

/dev/null is created by the kernel at boot time, because the kernel requires it. It's not udev specific.


1

The question is answered by myself now! The problem was that the crypttab entry for the second container was invalid. Even though I double-checked, I missed the error, and the update-initramfs didn't complain either. What do I take away from this? Always triple- or quadruple-check such critical things, as it can often save you a lot of hassle (and others ...


1

service is now handled by systemd, and systemd is quiet when everything goes well. There doesn't seem to be a way to change that. Lennart added a TODO item to add a verbose mode restoring behaviour similar to what you're after, but three years later it's still in the TODO list! You could always define a shell function to do something like service tor ...


1

Custom services should stay in /etc/systemd/system. Since you are doing forking, what I would do is use the PIDFile= directive. When the service is stopped, it will delete the file. However, on startup, it will not write to the file. It's up to your service to write to it. The recommended place to have it write to is /run.


1

There is no magic in such a case. You can't use a Debian package's static copy to install both the systemd and sysV files because this will duplicate services. So, the solution I found was to deliver the configuration files as normal files inside the deb file, and then to implement a routine inside the postinst / postrm scripts that identifies which init ...



Only top voted, non community-wiki answers of a minimum length are eligible