We have a couple of devices deployed on a large institutional network (a hospital). The devices are in this weird state where they keep getting updates to our software (we actually see the build number being updated) and they actually can talk to our servers, but for some reason the dashboard shows them as last connected months ago and we can’t trigger a BalenaOS update.
We’ve talked to the network administrators at the institution and they said they’ve “reclassified” the old endpoints (vpn.resin.io and pubnub.com) and the devices should be able to talk to those endpoints. My questions are:
Are those endpoints still valid?
Is there a way to trigger a BalenaOS update remotely without using the dashboard?
Hi there, just enabled support access on device 61d63ccf5deb951bf84698d19646a783aee860c3ff28984e91f53be6397b2d. I’ve checked the latest requirements, and it seems like the people at the hospital have already updated their network firewall to support those.
@davidg I’ve had a look at the device and see what you mean; it’s running a 1 month old release but says it hasn’t been online for 5 months. Do you have any other devices on the same local network that you could also share support access to? In some cases we are able to use the working (online) device to connect locally to the offline one to further debug.
I’ve just granted access to 273d427480ee3f90a03a4187ff72ea89e756e8eb6e8ff87af9154e1b1e10e6. To be honest, I haven’t seen that device connect to our services in a while (about a month now), so not sure it’ll be accessible to you. Please let me know if you learn anything.
And yes, you are correct. Weirdly enough the device seems to be getting software updates just fine, so at least some part of connectivity to Balena must be working. My guess is that whatever long-running connection to Balena’s cloud needs to be up for the dashboard to issue commands and read logs is what is dead.
@davidg ahh they are both offline, I thought one might have still been OK. That means I won’t be able to connect. To answer your original question though, the old endpoints are indeed still valid so that wouldn’t be the cause of the issue.
Sometimes you can trigger a reboot of a device and bring it back online by toggling persistent logging on; however in the case of your devices they are running an older supervisor version that doesn’t support persistent logging so that may not work. If you wanted to try this you can toggle it in the device configuration (we can’t make changes to your devices otherwise I’d try it for you).
I’ll definitely try that! I’ll also ask the technicians at the hospital to take a look at the devices, maybe they’ve been disconnected for cleaning or something. Thanks so much for your help!