Update loop or failed due to upgrade_available

Hi,

After updating the board (Variscite imx8mp dart) the deployment status changed from ‘Downloading’ to ‘Rebooting’, the board goes into reboot and boot into the newly updated partition.
On the demo server side, after the reboot, if the ‘upgrade_available’ environment variable equal ‘1’ then the deployment status changed to ‘failure’. On the other hand, if ‘upgrade_available’ equal ‘0’, the deployment status changed back to ‘Downloading’ and the whole update process starts over again.
Yocto scarthgap
mender 4.0.6
mender-client 3.5.3

Any idea?

Thanks,
Yair

Hi @yairbenavraham,

Sorry, I can’t completely follow. So do I understand correctly:

  • you are not using the standalone flow in the described situation. The deployment is triggered through a server instance, such as hosted.mender.io.
  • the reboot occurs automatically, and exactly once
  • after that, the board is running the new software, but the deployment on the server is never finished and stuck in rebooting. Is that correct?

One thing that sounds strange is the versions:

  1. mender 4.0.6 AND
  2. mender-client 3.5.3
    Do you have both installed on the device? Can you eventually provide the list of installed packages? Two clients running at the same time would definitely have unexpected effects.

The handling of upgrade_available and bootcount should not require any manual handling. You can check their correct operation using the integration checklist

Greetz,
Josef

Hi @TheYoctoJester ,

Your understanding is correct: deployment is triggered through a server instance, reboot occurs automatically to the newly updated partition, deployment on the server is never finished: stuck in rebooting but after a while changed to downloading.
The installed and running mender is Mender 4.0.6.