Cannot stop container-management service

[sophokles73]

Describe the bug
I have run the latest 0.0.6 image for x86_64 and am not able to restart (nor stop) the container-management service in order to pick up some changes I made in some deployment manifest files.

To Reproduce
Steps to reproduce the behaviour:

1. start up leda image in QEMU
2. log in as root
3. wait for systemctl status container-management to indicate that the service is up and running
4. restart the service using systemctl restart container-management
5. wait forever ...

Expected behaviour
container-management service stops and restarts after a few seconds

Leda Version (please complete the following information):

• Version: 0.0.6 as downloaded from GitHub releases
• Machine: qemux86_64
• Connectivity: transparent internet access

[sophokles73]

I have killed the QEMU process and started up leda again. The changes I had made to the manifest files have been picked up and now I am also able to shutdown leda properly.

[mikehaller]

I have witnessed the same or a similar problem on Raspberry Pi: systemctl restart container-management would hang indefinitely. In that case, the device was under heavy load and running on a possibly broken SD-Card (CRC errors when trying to back it up).

We were using a Kuksa.VAL Seat Service example container reconfigured with physical CAN-Bus access, host networking and in privileged mode. It's unclear what the root cause was, but we could not properly shut down the system any more. The Kanto container-management socket interface was not coming up any more (hence, kantui and kanto-cm were not able to connect to the container-management daemon).

Also stopping the service did not properly work either.
The file /run/container-management/netns/default was locked, but unsure whether this is related.