Stratis Is Red Hat's Plan For Next-Gen Linux Storage Without Btrfs

[Pajn]

I've got it on about seven drives, and I've been using it since 2013 and never had a problem. The speed is fine too.

The sad thing is that partity raid still is not stable. With raid 5 on 3 disks I get the same security as btrfs raid 1 (not ordinary raid 1) but with more available space, and with four disks I raid 6 can give me double security for the same available space. Currently btrfs does not allow me to have two disks die without data loss (or sure, I guess I could configure raid 11 but multi level raid is always a PITA and parity is just so much nicer). But yeah, I will still use btrfs on my next NAS, even with just raid 1 it still beats mdadm + ext4 thanks to self replicating checksumming. And snapshotting and subvolumes are awesome features that makes great backup solutions super easy and fast.

[speculatrix]

I'm surprised Poettering hasn't created systemd-fs and subsumed the file system into systemd altogether!

[Elyotna]

Stratis is developed in Rust and Python.

kthxbye

[lordnaikon]

I wonder how this compares to the TFS Filesystem written for the Redox-Os (both in Rust, too) https://www.phoronix.com/scan.php?pa...ystem-Rust-ZFS and https://www.phoronix.com/scan.php?pa...-OS-2016-State

[gbcox]

This seems like an odd direction to me. It will take 1-2 years to release the new software that ties together the filesystem and services needed. Then what, 6-12 months to stabilize for production use? They're saying BTRFS won't be acceptable in that timeframe, even though it's used by Facebook for their infrastructure today?

BTRFS is used in a limited, targeted fashion at FB. Redhat internally has invested in XFS - they made that decision and aren't going to trust their customer production workloads on a system they aren't actively developing and don't have in-house expertise.

[bug77]

ext4 lacks many "modern" and advanced features like builtin compression, encryption, RAID, deduplication and more. Even it maintainer (Theodore Ts'o) said that btrfs is future for Linux storage.

Not surprising since ext4 is mostly ext2 with sugar on top (journaling). I.e. it's older than people think it is.
It's great for home use, because it has the widest tool support though.

[onicsis]

I'm surprised Poettering hasn't created systemd-fs and subsumed the file system into systemd altogether!

Stratis 4.0 = Poettering-FS.
Already has too many moving parts, dependency and it breaks easily.
At least ZFS or BTRFS are self contained

[Pajn]

I wonder how this compares to the TFS Filesystem written for the Redox-Os (both in Rust, too) https://www.phoronix.com/scan.php?pa...ystem-Rust-ZFS and https://www.phoronix.com/scan.php?pa...-OS-2016-State

TFS is a file system, Stratis is a userspace deamon that will use XFS and device mapper (and maybe in the future LVM if redhat can harras the LVM devs to add features they need) and try to emulate snapshots, subvolumes, pools on top of that. Basically Stratis will only give you nicer CLIs, not anything you can't already do.

Seems to be no mentioning of checksumming or how they will deal with problems in all other userspace programs. I will use whatever that solves all problems first but I have a hard time seeing how Stratis can ever solve everything, or even enough things to outweigh the additional problems it will create.

[Zan Lynx]

Meanwhile 8 years later..... Btrfs is still slow as hell and buggy... If RH can pull this off in just 2 years, Fantastic.

I've been running btrfs on my home fileserver since 2012. I started with two 2 TB WD Reds. I've added four more drives and replaced the 2 TB drives after one failed. It's up to 30 TB total now in RAID10 configuration.

In all that time it hasn't failed me. It properly corrected all of the bad data when one drive went bad. And it has been great to be able to rebalance to new RAID configurations without shutting down and restoring from backup.

[blackiwid]

I disagree on this one.

https://www.maketecheasier.com/best-...ystem-for-ssd/

maybe the overhead of journaling is acceptable and performance is more important than longer livetimes of ssds, But you should consider than ssd will replace year by year more harddisks even in data centers.

So you reduce the life-time of all of them. A journaling filesystem for a ssd world I dont get it.

I guess for that data centers righer speed > lifetime of hardware. But for home users the opposite is likely true. I would say whatever its their enterprise distro. But they implement it in fedora, where it makes little sense.

Thats the bad side about fedoras model, its only the test place for enterprise features, if your needs differ from enterprise needs, nobody cares I mean they will not remove btrfs support, at least on packet level but the installer could stopp supporting it. that can be a problem for users. It would lead to strange hacks that you cant use the normal installer anymore or stuff like that and somehow install it another way.

For a distro that has a chance to become a top player in the desktop linux sector a problem. Then most users keep using their garbage ubuntu based distros.

Even apple has now a COW filesystem, the linux / freebsd community had a advantage we got that faster, now we/they through that away for some strange hack that never will become as good as such filesystems. I will hear soon the first people that say you have to install apple for better filesystems than linux has.