Generic System for Safe Rootfs and Kernel Upgrades Without Single-Point of Failure

Explore a comprehensive conference talk on implementing a generic system for safe rootfs and kernel upgrades without single points of failure in embedded systems. Learn about a minimal kexec/initramfs system that addresses challenges in fetching, verifying, and upgrading firmware while ensuring device resilience against power failures. Discover techniques for streamlining upgrades on legacy systems with limited space, reducing factory flashing time through parallelization, and safely managing kernel upgrades using drop-in files. Gain insights into unit testing, handling regressions during beta testing, and the speaker's experiences contributing to Linux drivers and preparing custom ARM boards.

Introduction
First solution
Second solution
What is danger
What we present
Refresh request
Boot order
Upgrade method
Upgrade process
Colonel upgrades
Questions
Unit Tests
Resources
Next Steps
Yocto
Unit Test
Key Rotation