Btrfs at Scale: How Meta Depends on Btrfs for Our Entire Infrastructure BY The Linux Foundation (May 25, 2023)
Posted by fenix0000000@reddit | linux | View on Reddit | 127 comments
Video (41 min): Btrfs at Scale: How Meta Depends on Btrfs for Our Entire Infrastructure - Josef Bacik, Meta
Explained by Phoronix :
Josef Bacik, a prominent Btrfs engineer at Meta, wrote about the magnitude of impact for Meta's Btrfs usage:
"The Meta infrastructure is built completely on btrfs and its features. We have saved billions of dollars in infrastructure costs with the features and robustness of btrfs."
With the scale to which Meta operates and their massive infrastructure, Btrfs is attributed as having saved "billions of dollars" thanks to its advanced feature set and robustness. An interesting anecdote for those that continue to question Btrfs or its suitability for use in production environments.
More commentary can be found via this LKML thread amid the ongoing discussion over Bcachefs in the mainline Linux kernel.
From: Josef Bacik <josef@toxicpanda.com>
To: Kent Overstreet <kent.overstreet@linux.dev>
Cc: "Aquinas Admin" <admin@aquinas.su>,
"Malte Schröder" <malte.schroeder@tnxip.de>,
"Linus Torvalds" <torvalds@linux-foundation.org>,
"Carl E. Thompson" <list-bcachefs@carlthompson.net>,
, ,
Subject:
Date: Sat, 9 Aug 2025 15:21:56 -0400
Message-ID: <20250809192156.GA1411279@fedora> ()
In-Reply-To: <>
On Sat, Aug 09, 2025 at 01:36:39PM -0400, Kent Overstreet wrote:
> On Thu, Aug 07, 2025 at 07:42:38PM +0700, Aquinas Admin wrote:
> > Generally, this drama is more like a kindergarten. I honestly don't understand
> > why there's such a reaction. It's a management issue, solely a management
> > issue. The fact is that there are plenty of administrative possibilities to
> > resolve this situation.
>
> Yes, this is accurate. I've been getting entirely too many emails from
> Linus about how pissed off everyone is, completely absent of details -
> or anything engineering related, for that matter. Lots of "you need to
> work with us better" - i.e. bend to demands - without being willing to
> put forth an argument that stands to scrutiny.
>
> This isn't high school, and it's not a popularity contest. This is
> engineering, and it's about engineering standards.
>
Exactly. Which is why the Meta infrastructure is built completely on btrfs and
its features. We have saved billions of dollars in infrastructure costs with the
features and robustness of btrfs.
Btrfs doesn't need me or anybody else wandering around screaming about how
everybody else sucks to gain users. The proof is in the pudding. If you read
anything that I've wrote in my commentary about other file systems you will find
nothing but praise and respect, because this is hard and we all make our
tradeoffs.
That courtesy has been extended to you in the past, and still extends to your
file system. Because I don't need to tear you down or your work down to make
myself feel good. And because I truly beleive you've done some great things with
bcachefs, things I wish we had had the foresight to do with btrfs.
I'm yet again having to respond to this silly childishness because people on the
outside do not have the context or historical knowledge to understand that they
should ignore every word that comes out of your mouth. If there are articles
written about these claims I want to make sure that they are not unchallenged
and thus viewed as if they are true or valid.
Emails like this are why nobody wants to work with you. Emails like this are why
I've been on literally dozens of email threads, side conversations, chat
threads, and in person discussions about what to do when we have exceedingly
toxic developers in our community.
Emails like this are exactly why we have to have a code of conduct.
Emails like this are why a majority of the community filters your emails to
/dev/null.
You alone with your toxic behavior have wasted a fair amount of mine and other
peoples time trying to figure out how do we exist in our place of work with
somebody who is bent on tearing down the community and the people who work in
it.
I have defended you in the past, I was hoping that the support, guidance, and
grace you've been afforded by so many people in this community would have
resulted in your behavior changing. I'm very sorry I was wrong, and I'm very
sorry if my support in anyway enabled the decision to merge your filesystem.
Because your behavior is unacceptable. This email is unacceptable. Everything
about your presence in this community has been a disruption and has ended up
with all of our jobs being harder.
You are not some paraih. You are not some victim. You are not some misunderstood
genius. Your behavior makes this community a worse place to work in. If you are
removed from this community it will soley be because you lack the ability to
learn and to grow as a person and take responsibility for your behavior.
If you are allowed to continue to be in this community that will be a travesty.
Thanks,
Joseflinux-bcachefs@vger.kernel.orglinux-fsdevel@vger.kernel.orglinux-kernel@vger.kernel.orgRe: [GIT PULL] bcachefs changes for 6.17[thread overview]raw3ik3h6hfm4v2y3rtpjshk5y4wlm5n366overw2lp72qk5izizw@k6vxp22uwnwa
vga42@reddit
Good for them. That doesn't still mean it's a great choice for a personal desktop.
the_abortionat0r@reddit
Sorry but what? How is protecting your data not a good home feature?
vga42@reddit
I suppose it's stable these days, but btrfs has been of the most data-destroying filesystem I have ever used.
jabjoe@reddit
You're missing out.
grub-btrfs, self healing RAID, incremental backups by pushing snapshots to remote machines and more.
ZFS offers a bit more (built-in encryption) but isn't in the kernel, and that brings problems / limits.
vga42@reddit
Yeah I know what it can do. I currently have need for none of those things on my desktops. I'm aware that some people could use them for desktop too, though.
On servers many of those could be useful though.
the_abortionat0r@reddit
"data being safe is bad!"
jabjoe@reddit
On servers they are even more useful...
AreYouOKAni@reddit
You don't need them, but they are there and do not cost anything to have.
protoxxhfhe@reddit
btrfs has his own self limit by not being able to do raid, and his multiple problems all over, never saw any problem on zfs, it's still ext4+lvm or zfs
the_abortionat0r@reddit
The exactly one problem with raid 5/6 is exactly that, one problem.
Get your shit together dude.
jabjoe@reddit
The problem is just with RAID5/6, there is a "write hole" problem. Btrfs RAID isn't the only one with this.
shadeland@reddit
Is RAID5/6 fixed yet?
I need parity storage. I've got terabytes of files (GoPro footage from 2,000+ skydivers, screen recordings for courseware, etc.). That's why I use ZFS now.
jabjoe@reddit
Nope. Don't know it's even being worked on. I think most of us are happy with RAID0,1 or 10. From what I skimmed, it not easy and other RAID implementations have the same "write hole" problem. To be honest, I think it should be disabled if isn't safe.
shadeland@reddit
That's a bummer. Having 8 drives, and getting 6 drives worth of storage out of them is too amazing of a feature to pass up, so ZFS it is.
jabjoe@reddit
If it was a big deal, it would be being looked at. No one bothering about it tells me it's not thought that important. RAID 10 is faster and easier to repair and simplier.
ZFS is out of tree, which is a deal breaker to me, so until it's either in tree or there is a stable ABI agreed, I'm not interested. Stable ABI for external components is a constraint never been allowed in Linux before. The ZFS's license is immutable now, as the Linux kernel's, so they will never be compatible.
Long run, I think this will mean ZFS will be replaced. It's had a massive head start but Btrfs is a good enough replacement for most stuff. Competition is good, so maybe it's good there has to be competition due to licensing.
shadeland@reddit
Right now the Achilles heel is parity storage, which is just too valuable.
RAID 10 requires twice the drives. Parity just requires one or two drives to provide protection. I don't need speed, specially with spinning disks (I'd never put anything that required speed on spinning disks).
I've got literally Terabytes of video and my timeline for storage is my lifespan, so I need parity storage and I need the checksumming and healing for bitrot (and I like the encrypted data sets).
ZFS is developed on Linux first. The license incompatibilities aren't all that limiting. If there's another file system that supports encrypted data sets (beyond luks), parity storage you can count on, and checksumming/healing then I'll take a look. But that ain't btrfs now.
jabjoe@reddit
Pretty much every out of tree modules I've had to deal with, failed at some point on some kernel (or X in the case of Nvidia) update. (Ok, sometimes it's an easy fix, you can see arguments for some function has changed, so can fix yourself without it turning into a project. Still not ideal.)
Filesystem is too important to have that nonsense with. So no deal.
You can RAID 5/6 if you put mdadm under btrfs, but I agree that's not ideal. It's what Synology do for with their NAS. They still claim self healing for btrfs, so not sure what they have done.
LUK'ing is OK for encryption.
RAID 10 does require double disking, but that also makes it quite a bit faster, so it's not clear cut worse for all user cases.
If ZFS is what you want, fine, but it's not for me. It's advantages are into the weeds now.
klyith@reddit
But not subvolumes or nodatacow, and iirc ZFS snapshots don't have quite the same feature set as btrfs.
Transparent FS encryption is pretty nice tho.
jabjoe@reddit
I've not used ZFS as the out of tree is a deal breaker for me, but encrypted is the feature I'd have from it. I use btrfs on LUKS when I want encryption and that means RAID can only be mdadm, which means not self healing.
Synthetic451@reddit
Why is that the case? Can't individual btrfs devices be encrypted and then added to a btrfs array?
jabjoe@reddit
Mmm OK that should just work. LUKS the disk and add the encrypted block device to the btrfs RAID. Interesting. So LUKS'ing the single disk of btrfs is normal in laptops, so isn't even that odd really.
580083351@reddit
btrfs is the filesystem in Valve's Steam Deck. It makes sense for their usage because they have A/B partitions as their setup is immutable and they know every Deck has a battery which basically functions as an UPS--they don't need to worry about sudden power outages under normal circumstances.
klyith@reddit
In anything other than raid5/6 mode, power outages are not a problem for btrfs.
mdedetrich@reddit
Its funny how someone as knowledgeable as Josef Bacik can also be so clueless. The reason why btrfs work at scale in meta is that meta only runs btrfs in very specific configurations and there servers are in data centers like Equinix where they, for example, have literal rooms full of diesel generators to guarantee 100% uptime which means that meta never has to deal with unintended cold shut down due to power going out.
btrfs has had countless complaints regarding baseline reliability from users that don't have billions in budget and engineering resources to make the problems go away. Hell recently basic functionality like mounting btrfs volumes broke, yet again, Fedora https://www.reddit.com/r/Fedora/comments/1md7uk6/comment/n5zhuxe/
the_abortionat0r@reddit
Thanks for letting us know you have a mental illness.
It's insane you think companies would choose a file system that wasn't reliable and you have a whole ass conspiracy in your head vs reality where the facts are clearly in BTRFS's favor.
WaitingForG2@reddit
You made 14 replies under this post. Please don't gaslight others.
xiaopewpew@reddit
It is quite interesting "technical people" in this community would identify with this comment. Google's entire backbone network were initially built with less reliable but cheaper devices on purpose to save on cost and Google developed lots of software based techniques to mitigate expected outages.
>It's insane you think companies would choose a file system that wasn't reliable
Meta doesnt need to pick the most reliable file system, it picked a file system it understands the failure modes well, it picked a file system it can influence and it designed software/hardware/processes around the failure modes to mitigate them. Understand Meta adopted the FS around 2014 as far as I can remember. The FS at the time was not reliable by modern standards.
The same way Meta didnt pick the "best" source control for its monolithic codebase, it picked Mercurial over git primarily because the company was able to influence the dev community in Mercurial better to build specific features to support its huge monolithic codebase.
lelddit97@reddit
go back to phoronix where nonsense is more welcome, kindly!!!!
lue3099@reddit
Wrong.
WaitingForG2@reddit
Btrfs might have saved billions of dollars and might have netted hundreds of millions of dollars donations to Linux Foundation, but-
Why should bcachefs dev care about it? I don't think he gets a dime from anyone big using his fs anyway. And it's not like "more popular = better" is good metric either(Windows 11 is better desktop OS by that logic, for example)
Anyway, looks like an attempt to bully him out of Linux Kernel development.
Business_Reindeer910@reddit
No, it's pointing out repugnant behavior since all the attempts so far have not stopped it.
WaitingForG2@reddit
It's bullying.
For example, from https://docs.kernel.org/process/code-of-conduct.html
Tell me how is using such language is okay in professional setting? But as you pointed out, Linus himself wants him gone, so it's okay. What else to do? Kick person for made up rule because Linus doesn't like him? Right, better to wait until he resigns himself.
jinks@reddit
It's bullying in the same way the allies "bullied" the Germans out of France...
Business_Reindeer910@reddit
No, he's not being kicked out! If i contributed something to the linux kernel and they remove it, I am not being "kicked out".
jaaval@reddit
If the language is ok depends on what was said or done previously. You seem to assume calls for his removal are not justified.
nevyn28@reddit
A good reason not to use btrfs. Weird post.
the_abortionat0r@reddit
It being solid is a reason not to use it? What?
nevyn28@reddit
meta obviously.
didn't mean to offend the fedora tweens.
afiefh@reddit
By that logic, Meta also uses Linux and very likely gnu utils. Good reason not to use these either?
nevyn28@reddit
Yes, also a good reason to not have a sense of humour.
the_abortionat0r@reddit
Sounds more like you're just broken.
nevyn28@reddit
meanwhile, your username.
the_abortionat0r@reddit
Oh good one, anything to prevent talking about facts related to the topic.
Not sure why BTRFS being so good hurts you
nevyn28@reddit
Should change your name to captain generic.
the_abortionat0r@reddit
I'm on Arch and twice your age but whatever.
It's pretty clear you're the one in an emotional spiral right now.
nevyn28@reddit
Twice my age? I guess that explains your senility
the_abortionat0r@reddit
So it's just nothing but ad home from you because you don't understand file systems?
Business_Reindeer910@reddit
This is it. This is what i've been waiting to see!
MdxBhmt@reddit
This hits hard, I wonder if OOP had to deal with similar behavior previously.
Business_Reindeer910@reddit
OOP?
MdxBhmt@reddit
"Original" OP (meant for Josef Bacik here)
Business_Reindeer910@reddit
so not object oriented programming.. lol :)
MdxBhmt@reddit
Hah! I should have thought that OOP has a more used acronym around here haha
Business_Reindeer910@reddit
I mostly only frequent technical subreddits and i've never seen your version of OOP that I can remember
MdxBhmt@reddit
I probably got expose to it on a cruise to hobbydrama or subredditdrama, it makes more sense in places that crosspost a lot.
firewi@reddit
Not to derail, but I’ve been using Btrfs since 2011 (via QuantaStor) and had some catastrophic incidents that I was able to recover from backup - but man, it was impressive back then. So impressive that other storage providers were more than happy to slap on their own copy-on-write into their filesystem and call it ZFS or whatever was popular at the time. But btrfs is still a monster of a filesystem that is light years ahead of anything else on the market.
The fact that Facebook embraced it and moved their whole business infrastructure into this is not just a testament to its magical like speed and efficiency, but it’s sheer order of magnatude performance above all other filesystems.
People forget how much impact just one person can have in a community. If I would have stayed at my position using btrfs in that organization, Steve from OSNexus and Chris at Oracle would have kept making the integration tighter and worked out the issues I would run into.
I was stoked when Facebook took on the mantle of chief tester and I know it will keep getting better. Sometimes I wish I would have stuck with it, but that’s how the dice roll.
mishrashutosh@reddit
wait, what? raw performance isn't one of the strong suits of copy-on-write filesystems (nor should it be as they do way more things than journaling systems). pretty sure xfs and ext4 are faster than btrfs for most tasks.
Business_Reindeer910@reddit
You could have written that as a top level comment, not sure what it has to do with what i said.
xiaopewpew@reddit
The fact the guy works for one of the most toxic company should make right from wrong preeeetty obvious but I guess not.
the_abortionat0r@reddit
Nice ad hom
Business_Reindeer910@reddit
If we applied that process, then almost nobody would be working on the linux kernel at all sadly :(
xiaopewpew@reddit
Noone said anything about “supporting kernel devs”. Most of us dont pay a cent to use linux and most of these guys (i think including Kent) aint paid. Would be nice for the community to not dogpile on the guy, maybe thats too much to ask.
Linux dev community’s code of conduct is a joke, Linus himself lives above it. He is a nerd jesus the community looks up to and apparently he can do/say no wrong. However personalities like himself aint rare. I have seen ceos like him and people on the receiving end of him, it is not pretty.
Business_Reindeer910@reddit
I do think Linus himself should be called out for some his behavior indeed. IN fact, he was, he even went on vacation for it and came back better. Maybe needs another vacation.
BinkReddit@reddit
The email you posted is from August 2025 and should probably be a must read for everyone in this community.
axonxorz@reddit
El bot
MdxBhmt@reddit
It's almost like it should be on its own thread, it's completely different focuses.
vagrantprodigy07@reddit
Josef sounds like an asshole. Also, I'll never trust btrfs with data again.
ilsubyeega@reddit
then which fs should we use?
Business_Reindeer910@reddit
BTRFS has been the default fs for both opensuse and fedora for years now..
580083351@reddit
Default for Valve's Steam Deck too.
DM_Me_Linux_Uptime@reddit
Is it? I sold mine off a while back, but at the time you needed to run scripts to convert root and microsd to btrfs.
https://gitlab.com/popsulfr/steamos-btrfs
580083351@reddit
I checked just now.. the two root / A,B (5 GB) OS partitions are btrfs. /var and /home are EXT4.
Seems like a good compromise.. the fancy stuff for the OS partition and stable EXT4 for userdata.
Business_Reindeer910@reddit
I didn't actually know that one
YKS_Gaming@reddit
and many fedora users have been bitten by btrfs sending them to emergency mode recently because of an unfortunate kernel regression.
AleBaba@reddit
That is a bad bug, sure, but it didn't lead to data corruption, right?
YKS_Gaming@reddit
it didn't lead to data corruption, but did lead to users without a spare thumb drive and computer scrambling to get their computers booting again.
Had ntfs or ext4 ever sent people's computer unbootable, en mass?
heroBrauni@reddit
This bug could easily be fixed from the emergency shell.
No USB Stick needed.
Business_Reindeer910@reddit
one off bugs are not the problem that is being alluded to with btrfs though.
vagrantprodigy07@reddit
Any FS that isn't fragile as fuck.
the_abortionat0r@reddit
I love seeing people who just make shit up.
Every metric has shown btrfs to be rock solid but you seem to have some mental block to facts
vagrantprodigy07@reddit
Every metric? You are the one making shit up now. Go actually test file systems in power loss and data loss scenarios, and get back to me.
the_abortionat0r@reddit
Lol don't have to ask I already have during overclock/undervolts tests.
Btrsfs doesn't have a power loss issue. If you knew even the SLIGHTEST bit of what you are talking about you'd know that rumor stems from raid5/6 where under certain conditions can have issues after a sudden power loss.
That has NEVER been an issue for any other mode.
Get your facts straight kid before letting us know how clueless you are
vagrantprodigy07@reddit
I'm not talking about rumors kid. I'm talking about actual testing. You sound like the clueless one to me.
ilsubyeega@reddit
Can you please list them? I have no experience of it
Z3t4@reddit
Zfs, more features than btrfs, rock solid.
the_abortionat0r@reddit
Do you actually think that? Can you simply add or remove disks from your zpool? No? Then maybe read up on the topic
Z3t4@reddit
Can you encrypt or use redundancy besides "RAID 1" on btrfs? Each fs has its caveats, maybe you should read more too.
vagrantprodigy07@reddit
EXT4 is typically the safe answer. XFS is also good at recovery after issues. If you are serious about it, build out a VM with a few file systems, and do some testing by simulating sudden power loss. If you get data loss after that, try to recover your files. If you can't, it's fragile. Last time I ran that test with BTRFS (after losing data IRL), it failed miserably.
StephenSRMMartin@reddit
Anecdata.
I've used btrfs since it was made available in the mainline kernel.
I've had *one* issue with it. It was resolvable with its tools.
In that time, I've had failing disks, failing ram, and lived in a house with entirely too many power outages. No other issues whatsoever. Btrfs detected that my files weren't matching the md5sums, which was due to the ram corruption; so it protected my data despite hardware failures. The dying drive? Trivial to pull everything from and check integrity.
I run it on hdd. ssd. sd card. No issues.
Meanwhile, I've had ext4 totally fuck up and corrupt files three times somehow.
dijkstras_revenge@reddit
When was the last time you tried it?
Just_Maintenance@reddit
I think Facebook will make up for your absence
nevyn28@reddit
Definitely, world full of morally bankrupt zombies.
prey169@reddit
Yeahhhh I've said it before and I'll say it again, if btrfs combusted for me years after it is supposed to be stable, then i cannot reliably trust it. Esp when besides for btrfs, I have yet to see the case in over 10 years of working with Linux personally or professionally a fs combust to a state that wasn't recoverable
In comparison, bcachefs with 2 different devices, both that have survived numerous hard power offs due to power dying either battery related or just the power dying for my desktop. Bcachefs is pretty damn rock solid
For comparison also, btrfs barely lasted 1 year. Bcachefs is over 1.5 years strong on my desktop and 400 days on my laptop. And obviously xfs and ext4 are rock solid as well, but bcachefs just has so many more features built in
Just my 2c tho
Fit_Flower_8982@reddit
This is how the myth gets fueled. You hear haters smearing that btrfs will eat your data (*throwing side-eyes at the toxic bcachefs dev), btrfs eventually breaks because disk/hardware failures and user errors inevitably happen, and more public blame is thrown at btrfs despite the lack of evidence.
prey169@reddit
Same SD card and steam deck is running fine with ext4. It wasn't hardware related
580083351@reddit
The deck's OS filesystems itself are btrfs though.
prey169@reddit
SD cards are defaulted to ext4 unless you modify the format + automount scripts
Synthetic451@reddit
I'd imagine a copy on write filesystem on an SD card would perform poorly and probably be bad for the longevity of the nand.
580083351@reddit
Good to know, I never used a SD on it.. to this day people still have problems with their SDs dying on the Deck.. maybe they are buying the wrong ones or there is something wrong with the hardware/os setup where SDs are concerned.
Fit_Flower_8982@reddit
It makes sense if you were using an SD card. These rely on very basic microcontrollers with minimal maintenance and resilience capabilities, they are very vulnerable to power outages. Even high-end models remain quite unreliable.
Btrfs has redundancy in its internal structures and checksums for both data and metadata, which makes it more effective against logical corruption compared to ext4; it also protects against other forms of silent corruption like bitrot.
prey169@reddit
I mean, I get it. In theory it should be better.
It just hasn't been in that use case for me and in other usecases for others.
Ext4 and xfs just seem to be more stable from what I've seen over 10 years so far
dijkstras_revenge@reddit
Were you running raid 5/6?
prey169@reddit
No - it was on an sd card w/ simple zstd compression. It was for the steam deck
this was a little over a year ago at this point - so remembering the exact bug/errors I ran into are probably hard to pin down
Good_gooner6942@reddit
As an ext4 user, SD cards are also shit, right? You asked for trouble.
prey169@reddit
Ext4 has been rock solid for 2 years on the same SD card, and on the same steam deck
klyith@reddit
The thing about ext4 is, you have no idea if some of your data is corrupt or not. If your SD card has a bunch of bad blocks right now and they're only hitting file data (not difficult since metadata is a small %) then the filesystem will work just fine. But the data is bad.
Btrfs is not like that, when data does not match checksums it stops with errors. And if you have enough errors it refuses to mount at all, which is so you can do recovery before it gets worse.
That said, I wouldn't use btrfs on a SD card. Because SD cards are notoriously unreliable, and I'd never use them for anything of importance in the first place. So if they have corrupt data, I don't care. A SD card in your steamdeck for games, in my rpi music box, they're not worth anything more than wipe and re-image if something bad happens. The data isn't worth protecting.
prey169@reddit
Agreed but I lost 220gb out of roughly 300 used. I wouldn't care if I lost like 1 file. Luckily in the end it didn't matter. The one game without cloud save I was playing (the quarry) was not damaged
tesfabpel@reddit
That's probably the problem.
I had one SD Card completely die with the Steam Deck precisely after a year (it was a Lexar SD Card made for "gaming", whatever that means, compatible also with Nintendo Switch; bought in July 2022).
I then bought an SD Card (SanDisk Extreme Pro UHS-I, Class 10, U3, V30) in July 2023 and it still works perfectly.
prey169@reddit
This was san disk and the same SC card is still being used with ext4
Otherwise, I would agree
KnowZeroX@reddit
are you sure it wasn't the sd card? sd cards themselves aren't the most reliable media with "relatively" high failure rates compared to ssds
btrfs is the default filesystem of fedora and opensuse, if there really was such a problem it wouldn't be where it is
StephenSRMMartin@reddit
I use it on an sd card with zstd compression for my raspberry pi. Zero issues in 4 years.
vagrantprodigy07@reddit
I can't speak to bcachefs, because I've never tried it, but, as you've said, btrfs isn't a reliable file system. I'll keep saying it no matter how many downvotes meta employees and their fanboys throw at me.
FryBoyter@reddit
Just make a claim and call people who have had different experiences fanboys. Do you really expect a different reaction?
Let's assume you're right. Why then have various distributions such as OpenSUSE and Fedora, as well as other projects such as Synology's NAS, been using btrfs as their standard file system for years? If btrfs were really that unreliable, a lot of users would be having problems. Why do we hear so little about this?
vagrantprodigy07@reddit
A lot of users do complain about btrfs. Unfortunately fanbois such as yourself downvote them to hell. Rather than bitching on here, and doing the same thing you are accusing me of, ie bashing people who had a different experience than you with a product, why not to actually stress test the fs in power and data loss scenarios?
It's so easy to tell how many of you have never done this. As a storage engineer, it's something I HAVE to do for my job. Having done it on btrfs, I can tell you, it's fragile.
prey169@reddit
maybe give bcachefs a try - its pretty rock solid rn. And its nice having features like *background* compression and also encryption without having to use LUKs all built in. They can use more users to prove to more people that its stable :)
as always tho - keep some sort of backups since it _is_ relatively new compared to ext4/xfs
vagrantprodigy07@reddit
I may do that, I've got a few drives that don't hold critical data, that could use compression. Currently I use zfs for that, but it couldn't hurt to move one over to another fs for testing.
the_abortionat0r@reddit
You literally just made that up.
EverythingsBroken82@reddit
and i use it on laptop, server and desktop and in VMs. And VMs often are just shutdown externally without properly powering off.. i _NEVER_ had an incident. for like more than 3 years.
and bcachefs has some niche situations where the performance is absolutely the worst. but most end users just do not utilize it enough to actually experience this.
primalbluewolf@reddit
With the anecdata here, I can only conclude it was something you've done wrong, as my btrfs installs are fine going on 5 years now across several servers, desktops and laptops.
No comment on bcachefs as a fs - If I was willing to put code from controversial figures on my systems, Id probably be running reiserfs.
StephenSRMMartin@reddit
Anecdata indeed.
I've used since it was made available in the mainline kernel.
I've had *one* issue with it. It was resolvable with its tools.
In that time, I've had failing disks, failing ram, and lived in a house with entirely too many power outages. No other issues whatsoever. Btrfs detected that my files weren't matching the md5sums, which was due to the ram corruption; so it protected my data despite hardware failures. The dying drive? Trivial to pull everything from and check integrity.
I run it on hdd. ssd. sd card. No issues.
Meanwhile, I've had ext4 totally fuck up and corrupt files three times somehow.
archontwo@reddit
I support Kent on patreon and have done for years because I like his vision for a new filesystem.
But I can help but agree with Joseph. If you want engagement wrt you work, it should stand on its own merits not on the shortcomings of any other filesystem.
It does not help that Kent keeps pushing that as if it was the selling point of bcachefs.
Honestly, I have already checked out of this drama as I very much use BTRFS every day for years and never had a problem with it.
Synthetic451@reddit
The sad part is that Kent doesn't seem to know how to de-escalate. Even if he makes valid points about some of Linux's development processes, he frequently goes down this rabbit hole of needlessly defending himself and the merits of bcachefs. I always just ask "why?" Now is not the time to do that.
RileyGuy1000@reddit
I'm gonna be the stickler and say that I really just don't care for btrfs.
This post should take some of it's own advice and not try to pump up btrfs with claims that meta (a famously creepy and now anti-diversity company) uses it in their data centers. I would be bored, but Meta of all people? That's not an incentive for me to use it.
This is gonna be one of those anecdotes you hear about someone having a terrible experience, but my experience with it really just hasn't been to much benefit. Yeah, having FS compression would rock pretty hard, and I like the snapshot features, but never in my life have I had a filesystem corrupt itself so quickly.
My SSD was idle on a usb -> sata converter and I accidentally bumped it, knocking the connector loose.
Absolutely blew the FS to smithereens and was virtual unrecoverable. I could've maybe recovered it, but the filesystem was chunky applesauce at that point and it was easier to suck up the bit of cache I was using that drive for and reformat it as ext4.
Until btrfs handles sudden hardware power outages a lot more gracefully and is up to the battle-tested standard that something like ext4 is, I'll likely be sticking with ext4 because it "just works", is much faster for handling my file workload (game with lots and lots and lots of cache files), is battle-tested and resilient to many of the common faults that happen in daily computer usage (sudden power-outs, system lockups, etc.)
john16384@reddit
For what exactly? Anything that is large is already compressed. I can categorize my files roughly in videos, audio files, images and miscellaneous files. Each of these take up an order of magnitude more space than the next (ie. videos 20 TB, audio 2 TB, etc)
Of those, only a part of the miscellaneous stuff is not compressed already. The rest is all compressed, and so FS compression will do nothing.
RileyGuy1000@reddit
Anything large is not guaranteed to be compressed. Large log files come to mind, as well as a myriad of uncompressed image data.
Even if a lot of the common files you use aren't very compressible, a large portion of the underlying system still is, and in my use case, FS compression would help significantly.
natermer@reddit
It is cool how you managed to make a post about a video about btrfs usage until Linux kernel mailing list drama about bcachefs.
SaltDeception@reddit
I’m not even upset. I love Linux mailing list drama. It’s my greatest wish that Netflix or Apple or someone makes a show based on this stuff. It could easily be the best reality show ever made, hands down. They wouldn’t even need to embellish the drama like they typically do.