We want to push our new architecture to devices on resin, and this may have some considerable change in delta. I have observed it taking hours to calculate delta and push the new changes on wired network. Is there a way to optimize this. Can anyone educate me more on how delta works?
Is it worth to do delta updates incase there is some major change in the code, like many (some 100’s) commits away from the current master.
Hello, deltas are generated on the server on-demand and can take a while, mostly depending on the size of the images involved. They are also only generated once, so subsequent requests are served immediately. On the device, deltas are far more resilient to bad network conditions than regular updates and combined with their smaller size can result in substantial cost savings in bandwidth usage on cellular.
Is your application currently running a single container? Are you moving to a multi-container architecture? I encourage you to reach out via a private support channel to share details about your application and fleet in order for us to advise you as best as we can.