Update stuck in 'rebooting' state on hosted mender

Hello all :smiley:

I tryed to add a state script to my Jetson TX2. Once I updated my device through hosted.mender the device reboot correctly and the update is installed but the device state on hosted mender is stuck at “rebooting”. What could cause such behaviour?

To add a state script I crated a dedicated recipe that inherit mender-state-scripts. The purpose of the script is to cleanup mender.conf files and correctly split up the confs between /etc/ and /data partition. The script works well and give the expected result.

What did I miss?

I Finally found out that my device couldn’t connect to the internet after reboot that is why. My DNS configuration was not properly done therefore the device couldn’t tell hosted.mender that it has correctly rebooted.

1 Like

Hi @perceval :wave:

I am curious, how did you figure it out? Logs, troubleshoot? SSH ?

Hey @oleorhagen,

I had still an ssh connexion on the device so I could run commands on it and investigate.
first I noticed that I had network routing by pinging an internet IP (like 1.1.1.1 or 8.8.8.8) but no name resolution because ping mender.io failed.
The process responsible of name resolution when running a systemd based distribution is systemd-resolved. But the service was stopped in a failed state.

By looking at the journald errors and trying to restart systemd-resolved service, I understood that the folders /var/volatile/log and /var/volatile/tmp where missing. I created them and restarted the service. Et voilà I could ping mender.io. and the device came back on hosted.mender.

1 Like