On Thu Oct 2, 2025 at 6:58 AM UTC, Stuart Henderson wrote: >> However, mounted async, the default > > not true. read the description of -o in mount(8). try mounting > with async and you'll see *way* more damage if you can't provide > clean power. > > btw it's not just OS-level filesystems; there's another translation > layer between the OS and physical storage (needed for wear balancing > etc) and other requirements to make sure flash cells are properly > written (https://sabrent.com/blogs/storage/power-loss touches on > this). > > could things be better than they are now? possibly. > > but what you are doing with killing power during heavy writes > is pretty much guaranteed to cause damage, possibly permanent, > to your drives.
Hi Stuart, I was mistaken. You are correct, the default is 'nosync', not 'async'. I have not tested with 'async' only 'nosync' and 'sync'. I have tested with 4K sector HDDs and SSDs. My latest testing was on 512 byte sector HDDs. It makes sense that SSDs would need some remaining power to stay in a consistent state. I am not sure how HDDs handle power loss, but I've never (knowingly) had them damaged from having power cut. There may be another way to test this that's safer, other than cutting power. I have had crashes also cause this type of corruption. Perhaps something like halt -n might work? -Henrich
