Hacker News Reader

Comment by fire | Hacker News Reader

oh neat, ReFS isn't dead

interested to see performance metrics when tech youtubers start messing around with it

Oh, I used this and the performance and deduplication surprised me greatly in ReFS. I was using it as a place to store backups of my systems. The on disk vs the file size was so great that I had to rethink my offsite storage solutions. There was a recent article on here from a researcher at Microsoft on file size and deduplication.

Deduplication is awesome, it's really unfortunate how dangerous it is with things available on Linux. I've had ZFS corrupt itself twice and not after disabling it, on the same hardware. And btrfs died on me once - I'm not 100% sure it's exactly due to that, but there's a limit how much I'm willing to spend time copying terabytes.

Was the btrfs failure with RAID? Compared to LVM/dmraid with other filesystems on top... BTRFS is remarkably easy to fault.

I can reliably break BTRFS RAID using the reset switch on my system. Others behave fine via journaling and whatever, same devices/kernels/RAID level

you mind speaking more about the kind of size differences you're seeing? also, do you have a link to that article by chance?

https://www.usenix.org/legacy/events/fast11/tech/full_papers...

I had 66TB folder that appeared to be 21TB on disk. Granted there was significant duplicate data as this was backups of multiple systems.

holy crap

Besides this use case, it has been supported on Windows Server configurations.

Reply on news.ycombinator.com