Bcachefs File-System Pull Request Submitted For Linux 6.5

[phoronix]

Phoronix: Bcachefs File-System Pull Request Submitted For Linux 6.5

Capping off an exciting first day of the Linux 6.5 merge window is a pull request seeking to land the long-in-development Bcachefs file-system into this next kernel version...

https://www.phoronix.com/news/Bcachefs-PR-Linux-6.5

[Vorpal]

Seems interesting. How does it compare to btrfs feature-wise and performance-wise? Assuming I don't use RAID. And that I have been perfectly happy with btrfs stability (so let's not get into the flame wars about that, just a quick objective and value neutral comparison is all I want).

[timofonic]

Thanks for the news!

I hope Bcachefs merging try and feedback gets followed here. It's a filesystem that differs from others, with quite interesting ideas and different approaches.

Btrfs, XFS and ZFS are interesting in many ways. But I would like the best of all them in a mainline filesystem and more. Who knows, maybe Bcachefs will bring that and someday make them and even EXT4 totally not just obsolete but redundant too.

[Developer12]

I'm overall pretty happy to see this land, though I would kinda prefer it if they did more to protect not just the data, but the metadata too.

They were talking on LWN about implementing an online fsck system like XFS, but all that really is to me is an online amputation system. Yeah, when you hit an error you may be able to make the filesystem "consistent" but that might not actually mean "undamaged."

This is a pretty stark contrast to how ZFS works, where you have redundant metadata structures and they're all individually checksummed so you can fall back to a good copy. You can actually restore it fully, not just say "this is damaged, remove it to become consistent again" or "we don't know where this folder goes anymore, move it to lost+found."

That's what people mean when they say it doesn't even make sense to run fsck on ZFS. It's not like there's an online fsck running in the background checking for consistency, it's that it's not needed because it's actually hard to do meaningful damage where repair is needed.

full disclosure: I have no idea how BTRFS handles metadata repair on a detailed level, so I can't comment there.

[Mitch]

I'm overall pretty happy to see this land, though I would kinda prefer it if they did more to protect not just the data, but the metadata too.

They were talking on LWN about implementing an online fsck system like XFS, but all that really is to me is an online amputation system. Yeah, when you hit an error you may be able to make the filesystem "consistent" but that might not actually mean "undamaged."

This is a pretty stark contrast to how ZFS works, where you have redundant metadata structures and they're all individually checksummed so you can fall back to a good copy. You can actually restore it fully, not just say "this is damaged, remove it to become consistent again" or "we don't know where this folder goes anymore, move it to lost+found."

That's what people mean when they say it doesn't even make sense to run fsck on ZFS. It's not like there's an online fsck running in the background checking for consistency, it's that it's not needed because it's actually hard to do meaningful damage where repair is needed.

full disclosure: I have no idea how BTRFS handles metadata repair on a detailed level, so I can't comment there.

Isn't metadata redundancy and checksumming a thing in bcachefs? I see references here.
https://bcachefs-docs.readthedocs.io/en/latest/options.html

Haven't yet tried bcachefs, but I'm planning on using it after it's in the kernel, more and more as it's baked in.

[Developer12]

Isn't metadata redundancy and checksumming a thing in bcachefs? I see references here.
https://bcachefs-docs.readthedocs.io/en/latest/options.html

Haven't yet tried bcachefs, but I'm planning on using it after it's in the kernel, more and more as it's baked in.

Ah you're right, there are "metadata_replicas" and "metadata_checksum" settings listed there.

Hopefully that gets employed to full effect when they implement online fsck? I kinda wonder what they're doing there now, if online fsck isn't implemented.

[Quackdoc]

super excited to be able to try this on my main machine and not something I use for off again on again usage

[vladpetric]

Praying that it might happen this time ...

[Quackdoc]

Praying that it might happen this time ...

I doubt it will, I assume this will be punted back and forth more then a couple times. if it makes it in before 6.5 I would consider that extremely lucky

[archkde]

I'm overall pretty happy to see this land, though I would kinda prefer it if they did more to protect not just the data, but the metadata too.

Metadata protection is kinda the standard for all except the worst (e.g. FAT) filesystems, and pretty much a prerequisite for data protection anyway (you won't be able to access your data if the metadata claims it's not there, after all). So it is safe to assume that any file system offers at least the same amount of protection for metadata as it does for data.