ROS systemd service terminating itself

Navio / Edge Raspbian
1

Hello,

I am trying to automatically start MAVROS on boot according to the tutorial (https://docs.emlid.com/navio2/common/dev/ros/).

Running roslaunch mavros apm.launch manages to launch without any issue. I created a systemd service using the exact same configuration as given, and enabled it on boot.

When powering on the drone it connects successfully but I realized that it is disconnecting / reconnecting continously after about 1.5 minutes. By removing the restart-on=failure I can see that it fails with the following log output:

â—Ź ros.service - ros
   Loaded: loaded (/etc/systemd/system/ros.service; enabled; vendor preset: enabled)
   Active: deactivating (final-sigterm) (Result: timeout)
Cntrl PID: 399 (bash)
   CGroup: /system.slice/ros.service
           └─499 /usr/bin/python /opt/ros/kinetic/bin/roslaunch mavros apm.launch

Oct 16 06:20:28 navio bash[399]: [ERROR] [1539670828.329754664]: FCU: PreArm: Throttle below Failsafe
Oct 16 06:20:58 navio bash[399]: [ERROR] [1539670858.324944037]: FCU: PreArm: RC Roll not configured
Oct 16 06:20:58 navio bash[399]: [ERROR] [1539670858.326239893]: FCU: PreArm: Compass not calibrated
Oct 16 06:20:58 navio bash[399]: [ERROR] [1539670858.327425704]: FCU: PreArm: Compass not calibrated
Oct 16 06:20:58 navio bash[399]: [ERROR] [1539670858.328863895]: FCU: PreArm: check firmware or FRAME_CLASS
Oct 16 06:20:58 navio bash[399]: [ERROR] [1539670858.330039811]: FCU: PreArm: Throttle below Failsafe
Oct 16 06:21:07 navio systemd[1]: ros.service: Start operation timed out. Terminating.
Oct 16 06:21:07 navio systemd[1]: Failed to start ros.
Oct 16 06:21:07 navio systemd[1]: ros.service: Unit entered failed state.
Oct 16 06:21:07 navio systemd[1]: ros.service: Failed with result 'timeout'.
Oct 16 06:21:07 navio bash[399]: process[mavros-2]: started with pid [549]
Oct 16 06:21:07 navio bash[399]: [mavros-2] killing on exit
Oct 16 06:21:07 navio bash[399]: [rosout-1] killing on exit
Oct 16 06:21:07 navio bash[399]: [master] killing on exit
Oct 16 06:21:07 navio bash[399]: shutting down processing monitor...
Oct 16 06:21:07 navio bash[399]: ... shutting down processing monitor complete
Oct 16 06:21:07 navio bash[399]: done
Oct 16 06:21:07 navio bash[399]: Unhandled exception in thread started by
Oct 16 06:21:07 navio bash[399]: sys.excepthook is missing
Oct 16 06:21:07 navio bash[399]: lost sys.stderr

Image and ardupilot version (from emlidtool):
Info

Vendor: Emlid Limited
Product: Navio 2
Issue: Emlid 2018-06-05 831f3b08594f2da17dccae980a2e3659115ef71f
Kernel: 4.14.34-emlid-v7+
RCIO firmware: 0xb09979ae

Ardupilot Info

On boot enabled: ['arducopter']
Arducopter - active (running)
Version: arducopter-3.5
Frame: arducopter

/etc/systemd/system/ros.service

[Unit]
Description=ros

[Service]
Type=forking
ExecStart=/bin/bash -c "source /opt/ros/kinetic/setup.bash; /usr/bin/python /opt/ros/kinetic/bin/roslaunch mavros apm.launch"

[Install]
WantedBy=multi-user.target

Any help greatly appreciated!

3

Hi @tobiasfriden,

Try to add TimeoutSec=infinity string in [Service].

4

@tatiana.andreeva thank you, this seems to have resolved the crashing.

However the service seems to never reach active state, after clearly starting everything it is still reporting Active: activating (start) since Wed 2018-10-17 06:38:59 UTC; 5min ago.

Is this expected?

5

@tobiasfriden,

Seems like forking type isn’t the right choice for this kind of service. Change it to simple and service will keep active (running).

6

This works fine, thanks! Maybe it should be changed in the documentation?

1 Like
7

@tobiasfriden,

Glad to hear it helps!

We’ll fix it in our docs in the nearest time. Thanks for noticing!

9

This topic was automatically closed 100 days after the last reply. New replies are no longer allowed.