Multicontainer project that works on RPI3 fails on Fin board (avahi & supervisor interaction)

I have recently encountered a similar pattern of installing starting installing stopping with a certain container: see Container in endless installing - starting - installing - stopping loop since upgrade from Resin Os 2.15.1+rev2 to balenaOS 2.27.0+rev1

Maybe your problem has the same root cause as mine. I have just reported my problem, so also now solution is yet known for that problem.

Hello, before I had the time to create a simpler project I tried a similar one that I found here: https://github.com/cape-io/signalk-server-balena

I pushed it on my fin and got the same problem. This time though the problem occurs in influxdb container instead of the signalk container, and the only pattern I see is that in both cases the problem occurs on the last container to be defined in the compose file.

Hi there,
We would need the device dashboard url for a device inspection.
What version of BalenaOS are you using?

Hello, do you mean this ?

https://dashboard.balena-cloud.com/devices/858fb1da2f687dfb15142b6841479a9d/summary

The device runs fincm3-2.26.0+rev1-dev-v8.0.0

I have enabled the public URL but as the container for port 80 is in the bounce loop it doesn’t work. I have granted support access for a week also.

Let me know if something else is required.

Thanks,

Alex

@krital Thanks, we’re having a look at it.
It looks like the container keeps restarting because of config changes and in the application you listed above, the server closes the process when there’s an error, so it might be for example a missing dependency in the Fin image (which would also explain the fact that it works on a Rpi3).
We’ll keep you posted

@krital Meanwhile, could you please verify that you’re not running into this issue?

Hi @thundron,
many thanks for this. I don’t have time to look in detail but ordering ports as ascending seems to stop the loop and my container starts! Will examine in more detail in the next few days…

Alex

2 Likes