Re: systemd service fails to run with "ConditionACPower=true was not met"

[Richard W.M. Jones]

On Fri, Aug 05, 2022 at 11:36:13AM +0100, Barry wrote:
>  
>  
>  > On 5 Aug 2022, at 10:48, Richard W.M. Jones <rjones(a)redhat.com&gt; wrote:
>  > 
>  > Fedora 36, plocate-1.1.15-3.fc36.x86_64, systemd-250.8-1.fc36.x86_64
>  > 
>  > /usr/lib/systemd/system/plocate-updatedb.service has
>  > ConditionACPower=true because they don't want updatedb to run
>  > when on a laptop battery.
>  > 
>  > Now, my machine is an Intel NUC which has an Intel mobile chipset but
>  > is very definitely not a laptop.  It doesn't have a battery and always
>  > runs off AC power.
>  > 
>  > But ...
>  > 
>  >  $ sudo systemctl status plocate-updatedb.service |& cat
>  >  ○ plocate-updatedb.service - Update the plocate database
>  >       Loaded: loaded (/usr/lib/systemd/system/plocate-updatedb.service; static)
>  >       Active: inactive (dead)
>  >  TriggeredBy: ● plocate-updatedb.timer
>  >    Condition: start condition failed at Fri 2022-08-05 10:29:57 BST; 5min ago
>  >               └─ ConditionACPower=true was not met
>  > 
>  >  Aug 03 14:59:37 dev5 systemd[1]: plocate-updatedb.service - Update the plocate
> database was skipped because of a failed condition check (ConditionACPower=true).
>  >  Aug 04 09:29:21 dev5 systemd[1]: plocate-updatedb.service - Update the plocate
> database was skipped because of a failed condition check (ConditionACPower=true).
>  >  Aug 05 10:29:35 dev5 systemd[1]: plocate-updatedb.service - Update the plocate
> database was skipped because of a failed condition check (ConditionACPower=true).
>  >  Aug 05 10:29:57 dev5 systemd[1]: plocate-updatedb.service - Update the plocate
> database was skipped because of a failed condition check (ConditionACPower=true).
>  > 
>  > Looking at systemd sources it seems as if ConditionACPower is checked
>  > here:
>  > 
>  > 
> https://github.com/systemd/systemd/blob/14e7bc2e77d2699498a1f74d7e4f905f1...
>  > 
>  > It's a bit hard to follow what the code is doing, but I think it's
>  > looking in /sys for power_supply devices.  This hardware has two, both
>  > with type == USB.  (I believe these are the two USB ports on the front
>  > panel that can provide power to other devices, they cannot power the
>  > machine itself.)
>  > 
>  >  $ ll /sys/devices/platform/USBC000:00/power_supply
>  >  total 0
>  >  drwxr-xr-x. 5 root root 0 Aug  3 09:00 ucsi-source-psy-USBC000:001
>  >  drwxr-xr-x. 5 root root 0 Aug  3 09:00 ucsi-source-psy-USBC000:002
>  >  $ cat /sys/devices/platform/USBC000:00/power_supply/*/type
>  >  USB
>  >  USB
>  > 
>  >  $ cat
> /sys/devices/platform/USBC000\:00/power_supply/ucsi-source-psy-USBC000\:00*/device/typec/port*/power_role
>
>  >  [sink]
>  >  [sink]
>  >  [sink]
>  >  [sink]
>  > 
>  >  $ cat
> /sys/devices/platform/USBC000\:00/power_supply/ucsi-source-psy-USBC000\:00*/online
>  >  0
>  >  0
>  > 
>  > Anyway, I think this is a systemd bug, right?
>  
>  That code has lots of debug logs. Try running systemd with log level debug and see which
> of those logs is printed. 
Good idea ...

# systemd-analyze set-log-level debug
# systemctl start plocate-updatedb.service 
# journalctl

The relevant section is attached in full.

It seems as if the code only finds the USB ports, they are all in
"sink" mode, it doesn't find any other power_supply devices so it
assumes battery?  There seem to be at least two things wrong here -
the USB ports are not "sinks" and there are no other power supply
devices.

I think this is a systemd bug.

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
libguestfs lets you edit virtual machines.  Supports shell scripting,
bindings from many languages.  http://libguestfs.org