Bcachefs Hopes To Remove "EXPERIMENTAL" Flag In The Next Year

[JMB9]

"You meant XFS in the Stratis part. Red Hat fucking hates OpenZFS because it makes Stratis look like a bad joke."

Yes, a typo - thanks for the correction: SGI is not Sun ...

But while XFS is a FS which is ok, Startis does look like a bad joke ... but Red Hat will handle it gracefully,
as IBM always does: ... "Nobody ever gets fired for buying IBM" ...

[Gabbb]

ReiserFS?

That's the only one example, though admittedly there are some parallels between these two projects.

[Vistaus]

tell me one filesystem in the linux kernel that had 10 years of deployment in enterprise production before they removed the experimental tag.

That too. Besides, where does it say that it's going to be declared stable? Removing the experimental flag is not the same as calling it stable. It's more like "we declare it good enough for daily use".

[User29]

In order for a filesystem to declare itself stable, 10 years of deployment in enterprise production is needed, possibly in varied scenarios, and possibly also with the occurrence of sudden incidents, even serious ones

This is a catch 22, isn't it?

[Weasel]

Yeah, yeah, we get it that you're simping hard for muh ZFS, but dial it down — not all of us share this level of fanboiness

At least it's production ready, unlike shills for Kent's toy.

[stormcrow]

I think a lot of people have a skewed idea of what 'stable' means in computer science. It doesn't mean "ain't gonna crash' or 'won't lose data'. It refers to API churn. Something is 'stable' and not experimental once the development churn calms down enough that the interface definitions no longer change rapid fire. What primarily remains is mopping up bugsplat from project scope growth. This is why you can alternatively consider the Linux kernel itself in its entirety as both "stable" because the user space interfaces rarely change, and not-at-all stable because the internal interfaces are in a constant state of flux and churn. It's that latter part that tends towards making the kernel less reliable than it might be otherwise because every time you make functional changes, there's a non-zero chance of introducing bugs (and why the distros maintain their own trees based on their own risk policies)

[sdack]

The article's title is somewhat misleading. No where did I see the word "hope" being used by the dev nor does it read very hopeful or ambitious in regards to the status. It rather reads very sober and realistic. If anything does it read like a reasonable response to someone who might be hoping to see the flag removed. It is a list of things that need doing before one can begin to hope or express such an ambition. Unless some parts have been left out is it only misleading to suggest the dev was motivated by hope or ambition.

[bkdwt]

Imagine running a 100TB+ storage with a experimental filesystem...

[stormcrow]

Imagine running a 100TB+ storage with a experimental filesystem...

Since that's only 5 20TB or 4 25TB spindles I'd call that a typical low end business storage test setup, especially in RAID 5 configuration.

[varikonniemi]

You meant XFS in the Stratis part. Red Hat fucking hates OpenZFS because it makes Stratis look like a bad joke.

Bcachefs still has a lot of work to do to be anywhere near the robustness and usefulness of OpenZFS.

openzfs recently went on and silently corrupted data for quite some time. Such has not happened yet on bcachefs.