We have a Raspberry Pi Zero W deployed at a client site that is locked and cannot be updated, log keeps printing the following
Killing application ‘registry2 .resin.io/…’
Failed to update application ‘registry2 .resin.io/…’ due to ‘[object Object]’
Restarting/rebooting the device doesn’t help. We also tried just about anything with the Supervisor remote API, again no luck. The device is running Supervisor version 6.1.2 so I suspect it may be related to this problem.
It’s almost impossible to go to the client site and replace the SD image, also even if we do this the same error may appear again.
We’re a IoT startup that has just launched our pilot programme, with live installations in a couple of different clients so having access to our devices is really important at this stage. If all goes well, we’d like to use resin.io for all our device fleet.
Is there anything that you can do remotely to help us get past this error?
I’d appreciate this a lot.
Hi everybody. We have released version 2.7.8+rev1 for the raspberrypi 1, raspberry pi zero and raspberrypi zero wireless.
This release should fix the above mentioned problem. Can you update to that version and let us know if you still see this issue?
Hi @hdemeillon,
Could you send me the dashboard link to the device and open the device for support access in order to further investigate your issue? This can be done through the dashboard, in the actions panel of the device.
Hi @hdemeillon the device is fixed up now. Your device was already running the new system, but it was running an old supervisor, which is known not to work on the new version, thus the OS version/application/etc… was not updated in the dashboard (the supervisor does that)
So here’s what happened:
the device’s OS was successfully updated; the supervisor update stage did not succeed for some reason (ie. the new supervisor for the device couldn’t be marked in the API), probably some transient API issue. In the update it is a warning currenctly, not a failure, so the updater progressed and rebooted
after reboot, the supervisor did not start as supervisor v6.1.2 does not work on resinOS 2.9.x, hence the stall
the fix was manually tagging the device for a supervisor update and running it
Most of the info was gathered from the resinHUP logs (available if you connect to the host OS in the web terminal or resin-cli), at /mnt/data/resinhup.