BCacheFS is being disabled in the openSUSE kernels 6.17+

https://lwn.net/ml/all/9032de2a-03a7-4f9e-9c8a-8bd81c5d1fc5@suse.cz/

by 6581 • 10 months ago

70 121 10 months ago

bgwalter• 10 months ago

[deleted wrongthink]

graemep2• 10 months ago

There is an apology for that comment and a rewording further down the thread. Evidently made by someone who is not a native speaker who did not realise how it comes across.

teekert• 10 months ago

Good addition,thanx.

I've been in a similar situation, letting everyone know I was fired. Apparently in the US this has a negative connotation, and they use "being let go" (or something confusing as "handing in/being handed your 2 weeks notice", a concept completely unknown here). Here we only have one word for "your company terminating your employment", and there is no negative connotation associated with it. This can be difficult for non-natives. We can come across very weird or less intelligent.

dbdr• 10 months ago

Funnily enough the apology ends with:

If the above offended anyone, I sincerely apology them.

Unless this was tongue-in-cheek, this kind of proves the point that language was the cause. The apology is a good move in any case.

t51923712• 10 months ago

Why would the "behave" comment mean anything different in Czech than in English?

The revised version, "Once the bcachefs maintainer conforms to the agreed process and the code is maintained upstream again" is still lecturing and piling on, as the LWN comments say:

https://lwn.net/Articles/1037496/

It is the classic case of CoC people and their inner circle denouncing someone, and subsequently the entire Internet keeps piling on the target.

zenoprax2• 10 months ago

"behave" in this context can refer to simply respecting existing norms about RC code freezing.

motorest5• 10 months ago

Ultimately that's the right call, and the inevitable one as well.

lupusreal• 10 months ago

The way the BCacheFS situation has been playing out is a tragedy. I had very high hopes for it.

InsideOutSanta2• 10 months ago

Yeah, this all seems so unnecessary. I hope Kent can either figure out how to work in the context of a larger team or find somebody who can do it on his behalf.

johnisgood4• 10 months ago

Once the BCacheFS maintainer behaves [...]

So, there are still behavioral issues here I take it? That is a bummer. This is not news to me, but I thought the situation has changed ever since.

Voultapher• 10 months ago

Don't hold your breath.

I've approached Kent two weeks ago and offered to help upstream changes into the kernel so he wouldn't have to interact with any of the kernel people ... he claimed without being able to explain why that having a go-between couldn't possibly help and that if he couldn't dictate the kernel release schedule he wouldn't want to ship the software anyway. And then proceeded to dunk on btrfs.

johnisgood4• 10 months ago

Same. I liked many of its features (actually, all features, see https://bcachefs.org) and I was waiting for it to become usable, but I guess that day will never come now?

So, the alternative is ZFS only, maybe HAMMER2. HAMMER2 does not look too bad either, except you need DragonflyBSD for that.

ThatPlayer2• 10 months ago

It's not unusable, I use it on a spare computer for fun, cuz I want tiering of SSD + HDDs. And this doesn't mean development has stopped, just not done in the kernel.

ahartmetz3• 10 months ago

What I expect to happen is that bcachefs stabilizes outside of mainline, and after that, it can be merged back because no large patches = not much drama potential.

qalmakka2• 10 months ago

RIP BCacheFS. I was hopeful I could finally have a modern filesystem in Linux mainlined (I don't trust Btrfs anymore), but I guess I'll keep on having to install ZFS for the foreseeable future I guess.

As I predicted, out of tree bcachefs is basically dead on arrival - everybody interested is already on ZFS, btrfs is still around only because ZFS can't be mainlined basically

Ygg2• 10 months ago

Wait. You don't trust Btrfs but you would trust BCacheFS, that's obviously very experimental?

rurban2• 10 months ago

Still more stable than btrfs. btrfs is also dead slow

phire• 10 months ago

Btrfs claims to be stable. IMO, it's not.

It's generally fine if you stay on the happy path. It will work for 99% of people. But if you fall off that happy path, bad things might happen and nobody is surprised. In my personal experience, nobody associated with the project seems to trust a btrfs filesystem that fell off the happy path, and they strongly recommend you delete it and start from scratch. I was horrified to discover that they don't trust fsck to actually fix a btrfs filesystem into a canonical state.

BCacheFS had the massive advantage that it knew it was experimental and embraced it. It took measures to keep data integrity despite the chaos, generally seems to be a better design and has a more trustworthy fsck.

It's not that I'd trust BCacheFS, it's still not quite there (even ignoring project management issues). But my trust for Btrfs is just so much lower.

kiney• 10 months ago

btrfs has many technical advantages over zfs

debazel• 10 months ago

Yes, like destroying itself and losing all data.

crest• 10 months ago

Can you give an example because to me it always appeared as NIH copy-cat fs?

sureglymop• 10 months ago

I've never had any issues with either ZFS or Btrfs after 2020. I wonder what you all are doing to have such issues with them.

jamesnorden• 10 months ago

Ah yes, the famous "holding it wrong".

pizza2342• 10 months ago

Just a few days ago I've had a checksum mismatch on a RAID-1 setup, on the metadata in both devices, which is very confusing.

Over the last one or two years I've experienced twice a checksum mismatch on the file storing the memory of a VMWare Workstation virtual machine.

Both are very likely bugs in Btrfs, and it's very unlikely that have been caused by the user (me).

In the relatively far past (around 5 years ago), I've had the system (root being on Btrfs) turning unbootable for no obvious reason, a couple of times.

pantalaimon2• 10 months ago

One lovely experience I had when trying to remove a failing disk from my array was that the `btrfs device remove` failed with an I/O error - because the device was failing.

I then had to manually delete the file with the I/O error (for which I had to resovle the inode number it barfed into dmesg) and try again - until the next I/O error.

(I'm still not sure if the disk was really failing. I did a full wipe afterwards and a full read to /dev/null and experienced no errors - might have just been the meta-data that was messed up)

Volundr• 10 months ago

Pre-2020, but I had a BTRFS filesystem with over 40% free space start failing on all writes, including deletes, with a "no space left on device error". Took the main storage array for our company offline for over a day while we struggled to figure out wtf was going on. Supposedly this is better now, but basically BTRFS marks blocks as data or metadata, and once marked a block will be reassigned (without a rebalance). Supposedly this is better now, but this was after it had been stable for a few years. After that and some smaller foot guns, I'll never willingly run BTRFS on a critical system.

patrakov• 10 months ago

I still have a btrfs with a big problem: more disk space used than expected. The explanation was helpfully provided by btdu:

    Despite not being directly used, these blocks are kept (and cannot be reused) because another part of the extent they belong to is actually used by files.
    
    This can happen if a large file is written in one go, and then later one block is overwritten - btrfs may keep the old extent which still contains the old copy of the overwritten block.

StopDisinfo9104• 10 months ago

btrfs is still around only because ZFS can't be mainlined basically

ZFS is extremely annoying with the way it does extend and the fact that you can’t mismatch drive size. It’s not a panacea. There clearly is space for an improved design.

nubinetwork2• 10 months ago

Underprivision your disks, then you don't have to worry about those edge cases...

cyphar• 10 months ago

This is being worked on (they call it AnyRaid), the work is being sponsored by HexOS[1].

[1] https://hexos.com/blog/introducing-zfs-anyraid-sponsored-by-...

accelbred• 10 months ago

I switched to ZFS for a while but had to switch back because of how much was broken. Overlayfs had issues, reflinks didn't work, etc. Linux-specific stuff that just works on kernel filesystems was missing or buggy. I saw later they added support for some of the missing features but they had data corruption issues. Also I doubt it'll ever support fs-verity.

I don't plan on giving ZFS or other filesystems not designed for Linux another go.

EspadaV9• 10 months ago

Might be harder to keep running ZFS on Linux after 6.18

https://www.phoronix.com/news/Linux-6.18-write-cache-pages

qalmakka2• 10 months ago

Killing ZFS on Linux would basically make Linux unsuitable for lots of usecases. What would you use instead? Btrfs, which keeps having stupid data corruption issues? Bcachefs, which is not yet stable and now it's being struck out of the kernel? LVM2+thin provisioning, which will happily eat your data if your data overlap? I hope some industrial players will force the kernel to drop this nonsense.

Heck no native filesystem besides btrfs has compression, I'm saving HUNDREDS of GB with zstd compression on my machines with basically zero overhead

koverstreet13• 10 months ago

The community is still growing (developers too!), and people have been jumping in to help out with getting DKMS support into the distros.

bcachefs isn't going away.

The SuSE guy also reversed himself after I asked; Debian too, so we have time to get the DKMS packages out.

the_duke• 10 months ago

This is a tragedy, bcachefs has so many great features...

rurban2• 10 months ago

Once the BCacheFS maintainer behaves and the code is maintained upstream again, we will re-enable... (As IMO, it is a useful feature.)

How cynical. It's the kernel maintainer, not the bcachefs maintainer, who does not behave and has a huge history of unprofessional behavior for decades.

happymellon4• 10 months ago

The bcachefs maintainer has added new features during bugfix windows, and lied about it.

pantalaimon2• 10 months ago

It's still an experimental module, the feature was about gathering more debug information.

nicman23• 10 months ago

How cynical. It's the bcachefs maintainer, not the kernel maintainer, who does not behave and has a huge history of unprofessional behavior for decades.

it is not like he was not explicitly warned.

boricj• 10 months ago

The original author later sent an apology email explaining that it sounded too harsh in English and it wasn't meant to be offensive:

https://lwn.net/ml/all/bece61a0-b818-4d59-b340-860e94080f0d@...

koverstreet13• 10 months ago

He also reversed himself when I asked him not to pull the rug out from users and explained that there is a plan for continued support.

The ever escalating drama and cynicism in the reactions this stuff gets though... bloody hell, what is with people these days?

self_awareness• 10 months ago

The reason is that people like Linus, because he's entertaining. And people don't like Kent, because he opposed Linus, who is liked. That's all there is too it. Like in some high school.

flykespice• 10 months ago

Do people.. find entertaining a Boss verbally abusing, with personal attacks, their co-workers?

Human nature is wicked.

fj23Z741GAh• 10 months ago

There was a time when Linus encouraged critics of "unprofessional behavior" to snap back at him:

https://lkml.org/lkml/2013/7/15/374

That is a reasonable compromise. Except when someone actually snaps back at him.

jtickle• 10 months ago

All of the "btrfs eats your data" bugs have been fixed and the people who constantly repeat them are people who relied on an experimental filesystem for files they cared not to lose. FUD all around. I have a btrfs on my home file server that's been running just fine for almost 10 years now and has survived the initial underlying hard drives mechanical death. Since then I have used it in plenty of production environments.

Don't do RAID 5. Just don't. That's not just a btrfs shortcoming. I lost a hardware RAID 5 due to "puncture" which would have been fascinating to learn about if it hadn't happened to a production database. It's an academically interesting concept but it is too dangerous especially with how large drives are now, if you're buying three, buy four instead. RAID 10 is much safer especially for software RAID.

Stop parroting lies about btrfs. Since it became marked stable, it has been a reliable, trustworthy, performant filesystem.

But as much as I trust it I also have backups because if you love your data, it's your own fault if you don't back it up and regularly verify the backups.

arccy• 10 months ago

"performant", it's still slow if you actually use any of the advanced features like copy on write.

FirmwareBurner• 10 months ago

Every CoW filesystem is just as slow. There's no magic pill to fix performance but it's a known tradeoff.

betaby3• 10 months ago

FUD all around

????

Don't do RAID 5.

Ah, OK, so not FUD

Stop parroting lies about btrfs.

I seee

plqbfbv• 10 months ago

All of the "btrfs eats your data" bugs have been fixed ... I have a btrfs on my home file server that's been running just fine for almost 10 years now and has survived the initial underlying hard drives mechanical death

In the last 10 years, btrfs:

1. Blew up three times on two unrelated systems due to internal bugs (one a desktop, one a server). Very few people were/are aware of the remount-only-once-in-degraded "FEATURE" where if a filesystem crashed, you could mount with -odegraded exactly only once, then the superblock would completely prevent mounting (error: invalid superblock). I'm not sure whether that's still the case or whether it got fixed (I hope so). By the way, these were on RAID1 arrays with 2 identical disks with metadata=dup and data=dup, so the filesystem was definitely mountable and usable. It basically killed the usecase of RAID1 for availability reasons. ZFS has allowed me to perform live data migrations while missing one or two disks across many reboots.

2. Developers merged patches to mainline, later released to stable, that completely broke discard=async (or something similar) which was a supported mount option from the manpages. My desktop SSD basically ate itself, had to restore from backups. IIRC the bug/mailing list discussions I found out later were along the lines of "nobody should be using it", so no impact.

3. Had (maybe still has - haven't checked) a bug where if you fill the whole disk, and then remove data, you can't rebalance, because the filesystem sees it has no more space available (all chunks are allocated). The trick I figured out was to shrink the filesystem to force data relocation, then re-expand it, then balance. It was ~5 years ago and I even wrote a blog post about it.

4. Quota tracking when using docker subvolumes is basically unusable due to the btrfs-cleaner "background" task (imagine VSCode + DevContainers taking 3m on a modern SSD to cleanup 1 big docker container). This is on 6.16.

5. Hit a random bug just 3 days ago on 6.16, where I was doing periodic rebalancing and removing a docker subvolume. 200+ lines of logs in dmesg, filesystem "corrupted" and remounted read-only. I was already sweating, not wanting to spend hours restoring from backups, but unexpectedly the filesystem mounted correctly after reboot. (first pleasant experience in years)

ZFS in 10y+ has basically only failed me when I had bad non-ECC RAM, period. Unfortunately I want the latest features for graphics etc on my desktop and ZFS being out of tree is a no-go. I also like to keep the same filesystem on desktop and server, so I can troubleshoot locally if required. So now I'm still on btrfs, but I was really banking on bcachefs.

Oh well, at least I won't have to wait >4 weeks for a version that I can compile with the latest stable kernel.

The only stable implementation is Synology's, the rest, even mainline stable, failed on me at least once in the last 10 years.

greyw• 10 months ago

Quota tracking when using docker subvolumes is basically unusable due to the btrfs-cleaner "background" task (imagine VSCode + DevContainers taking 3m on a modern SSD to cleanup 1 big docker container). This is on 6.16.

I had to disable quota tracking. It lags my whole desktop whenever that shit is running in the background. Makes it unusable on an interactive desktop.

yjftsjthsd-h5• 10 months ago

Yeah, no. I've had btrfs lose a root filesystem on a laptop with only one disk. No RAID, nothing fancy, well after it was supposed to be stable, on OpenSUSE where I assumed it would be well supported and pick good defaults.

Claiming that anyone reporting problems is lying is acting in bad faith and makes your argument weaker.

Also, "works for me" isn't terribly convincing.

M95D2• 10 months ago

I'm still waiting for an overlayfs that does read caching on the overlay without the need to format the backing storage.

rekoil• 10 months ago

Sounds like what bcache does? https://bcache.evilpiepirate.org/

This is what bcachefs is based on.

M95D2• 10 months ago

From kernel Documentation/bcache.txt:

You'll need make-bcache from the bcache-tools repository. Both the cache device and backing device must be formatted before use.

So, it's far from overlayfs. I could accept formatting the cache device, but not the backing storage.

duffyjp• 10 months ago

Very recently I setup a mergerfs mount for this. It’s very crude for my use case but works perfectly and I could use the existing volume as is.

I got partway thru setting up a script to copy recently accessed files from the HDD to the read-prioritized SSD.

My LLMs load up way faster, and I still have a source of truth volume in the huge HDD. It’s not something I’d use professionally though, way too janky.

betaby3• 10 months ago

Very theoretical question. If there was a usable microkernel, how hard would be it be to have an FS as a service? Are MacOS FS' processes or are they 'monolithic'?

yjftsjthsd-h5• 10 months ago

If there was a usable microkernel, then its filesystems would almost certainly be implemented in userspace; that's kinda the point of a microkernel. I can't speak to how Darwin does things, but I will point out that Linux and the BSDs have filesystems as userspace processes courtesy of FUSE; in their architecture that has a performance hit (caused by the context switching needed, I believe recently improved by use of io_uring), but it is a worked example of what you want.

betaby3• 10 months ago

I'm aware of FUSE and used NTFS driver though it back then. So we should expect similar performance with microkernels?

xenadu02• 10 months ago

So FSKit is public API and supports userspace filesystems but on macOS most built-in filesystems are kernel modules. On iOS everything except APFS is userspace.

Modern Darwin is mostly a monolithic kernel, though more and more things are moving to userspace where possible (eg DriverKit).

One interesting side-effect of various spectre mitigations is silicon improving the performance of context switches. That has the side-effect of decreasing the cost of kernel/userspace transitions. It isn't nearly as expensive as people still believe - though it isn't free either.

koverstreet13• 10 months ago

I think this would be the most usable microkernel these days, and yes, it does FS as a service: https://www.redox-os.org/

There's no inherent reason why a filesystem in userspace has to have a measurable performance impact over in-kernel, done right. It would be a lot of engineering work, though.

Context switches aren't that expensive when you're staying on the same core, it's IPIs (waking up a thread on another CPU) that's inherently expensive; and you need zero copy to/from the pagecache, which is tricky. But it totally could be done.

zoezoezoezoe• 10 months ago

dodged a bullet with this one. Migrated away from BCacheFS on my openSUSE deployments a few days ago because I could see the writing on the wall for a while.

BCacheFS is being disabled in the openSUSE kernels 6.17+

Related Stories