Guido U. Draheim

man fallocate(1) -p, --punch-hole Deallocates space (i.e., creates a hole) in the byte range starting at offset and continuing for length bytes. Within the specified range, partial filesystem blocks are...

That seems to be an undocumented feature, I have only found a hint on such a behaviour for the oneshoot-services. I did also see quite some code around where there...

ExecStopPost is meant to cleanup, so it may be wisest to not just let it fail. I'll moving a more detail stop-behaviour on to version 1.1 as the current systemctl.py...

note that a "systemctl.py stop service" will always execute ExecStop and ExecStopPost in the current implemention. In order to avoid that one would need to have a state model for...

I had been looking for the usage of "!!" lately, and I did not found any usage in a real service. In the past I had more success with just...

@EzraBrooks - I don't know if "systemd-timesyncd" can be run at all when systemctl.py sits on PID-1. That's my general assumption that systemd-xy tools should not be used in parallel...

On the develop branch, I had prepared a new parsing model for the Exec-prefixes during christmas season 2020, and I have done a first implementation of the nouser/noexpand flags just...

Tests look good, so (-, :, +) are implemented and (!,!!) are aliased to (+). Internally it is called "exe.check", "exe.noexpand" and "exe.nouser". To be in the next release.

@mvorisek - I'll check that next. The implementation was tested but I can not tell if it was skipped in that specific moment.

@mvorisek - the code looks as intended. I am not sure if I should try to recreate the environment that you have. Please "touch /var/log/systemctl.debug.log" and run your commands again....