Hi RAUC team, First of all, I'd like to thank the team for all the efforts and great jobs that have been done. RAUC is a great tool and your websites are a mine of useful information about embedded system update principles.
I'm currently evaluating RAUC to replace our current casync implementation. I've been trying the new Adaptive update method block-hash-index but I'm facing some issues : 1. since our rootfs images are built from scratch by our CI server, the file dates are always different. We don't to fake the date to a specific one (using BR2_REPRODUCIBLE). 2. As described in RAUC documentation and in your blog article, more chunks than only the required ones can be downloaded. due to the principle of squashfs block hash. 3. Some packages still includes the build date in their binaries. Our product is battery powered and use a Cat-M1 4G modem with a very expensive and limited data plain. So we need to download only the required files so we won't exceed the data plain and the machine can goes to sleep mode quicker. We can fix the issue #1 by using file checksum without using the date, as casync can do. We may fix the issue #3 by patching some packages. After I had a look to your (great) article : https://www.pengutronix.de/en/blog/2022-10-12-rauc-adaptive-updates.html I think the new tree-rsync-checksum adaptive method might be the fix for the issue #2. Here are my questions. Is there any public roadmap showing when it should be available ? Is there any public documentation explaining what would be the functionalities ? like ignoring the date when creating the checksums (like rsync/casync can do). Thanks for your answers. Regards. Alexandre Gambier Embedded Linux engineer IEM Group 109, chemin du Pont-du-Centenaire 1228 Plan-les-Ouates / Genève Tel +41 22 880 07 10 e-mail a.gamb...@iemgroup.com
