Currently, if the internet connectivity drops out, the download of a container is restarted and not resumed. This immensely prolongs the release rollout on devices with intermittent unstable connections.
Leandro Motta Barros: Hey folks, thanks for the suggestion!
balenaEngine actually resumes downloads in case of network connectivity issues (since Engine v19.03.23 / balenaOS v2.82.7). It will only give up if after about 20 minutes it was unable to download at least a single byte (in which case the download will indeed restart from zero later).
So, may I ask you to share a little more about the circumstances in which you are seeing this “container download restarted instead of resumed”? Do you believe your devices get disconnected for longer periods of time at once – such that our current resume policy is not working well enough? Or do you think it could be something else?
Thank you very much!
Leandro Motta Barros set the status to Completed
Actually, I am marking this item as completed to avoid bringing support or debugging discussions to our roadmap platform. As I said in my previous message, we are super interested in knowing the particulars of your case and to have the “resume pulls” feature tuned to work well. So, if it is not doing the job for your use case, please contact us through our support channels (Get in touch for support - balena). Thanks!
Chris Crocker-White: @Leandro should a container download also resume in the case of a system reboot? I noticed yesterday that I had a container downloading and rebooted the device expecting it to just resume but it didn’t, it started from the beginning again.