Does anyone even still use ext4 in this day and age? Have we finally transcended?

lmao bragging about just getting compression. Use zfs you mongoloid.

Topology has changed since, but damn son, don't dis zstd until you give it a try.
BTRFS is a worthy contender these days people meme it the same way they meme 'I tried linux seven years ago and it was shit, no thanks.'

ZFS still suffers from fragmentation when you kill the ARC, L2ARC is not persistent, no vdev removal.
It has Achilles heels.

Attached: Screenshot_20180405_004239.png (801x542, 71K)

Best filesystem for a single nvme ssd in a laptop? I am usually very close to the latest kernel, and I usually lvm encrypt the whole thing. Currently using ext4 because it's just the default. Also, any mounting options I should change in fstab?

Hold on, are you talking about desktop/laptop use?
Just use ext4 holy shit. The autists on this board I swear.

I have a btrfs-raid1-on-luks setup for my anime collection on two wd red drives, but for the system drive I use ext4-on-lvm-on-luks. Would be nice to move to btrfs there as well but btrfs does not allow swap files and creating btfs-on-lvm is too much of abstraction and is going to be slow.

Been using BSD UFS for 20+ years without any problems.

Attached: sto.png (409x207, 15K)

I use NTFS, the new technology file system.

>windows pleb
why do you exist?

You can use it on Linux as well.

I use Ext4 on everything because it works just fine. Even my RAID arrays.

>not using bcachefs and having your ssd be cache for your spinning rust

Can you disable the journal yet?

>Don't use nice things you autists
Why not just tell people buy a mac?

btrfs didn't 'just' get compression, it's had zlib and lzo compression for a long time. it just recently got /zstd/ compression
zstd is, if you don't know, a pretty recent algorithm made by facebook, with ratios around/slightly better than zlib, while being far, far faster (decompresses faster than lzo!)

I burn all my Blurays with NTFS

I use ext4 for my Debian install's /home and / partitions. I'm thinking of switching my /home to exFAT because I want to read and write from it from other OSes, namely from an old Windows 7 install i use to burn live CDs. Can anyone here attest to how good or bad exFAT is.

Attached: deviantart_402059477_Parsee sitting.jpg (1600x932, 1.5M)

ReiserFS is the only way to go

>not ReiserFS is killer

one job.

>ReiserFS slays the competition

neo-Jow Forums, a sight to behold
>i've heard about btrfs losing data, decade ago
>don't use superior technology
>doesn't know how to migrate ext4 on lvm to btrfs

ReiserFS takes out the trash (and buries it in the woods)

Btrfs is buggy. I've given up on it and waiting on bcachefs.

>>doesn't know how to migrate ext4 on lvm to btrfs
Does this solve its limitations with swapfiles? At least not yet.

Out of all Linux filesystems, XFS is the most sensitive to power loss and hard resets. My experience may be out of date (2010), but based on that experience I would only use XFS on a server with properly redundant power and only if I really needed the performance.

>not UDF

I use NTFS

What would I be using instead?

>using patreon-funded software

Using the latest and "greatest" mainstream software like Btrfs is the neo-Jow Forums thing. Paleo-Jow Forums hated complexity and churn almost as much as Uriel hated himself.

>Does anyone even still use ext4 in this day and age?
Yes, linux users who don't have autism.

ext4 is still heavily used. It's next to xfs one of the most stable and secure filesystems to this day, won't go away any time soon either.

Why do you need resizable swap partition? You're trying to find trouble where there's not any. Just create swap partition at same sectors as it is now and then migrate the rest of your fs to btrfs. If you *really* need resizable swap, you can shrink btrfs.
btrfs is far from mainstream and far from leatest, but it is greatest fs on linux for an average user. It's not stable enough for usage in data centres with thousands of PB of sensitive data that changes hw left and right, but it sure as hell is stable enough for a server/workstation/desktop use. It's much safer than ext4, it's better for SSDs, it offers fuckton of really useful features and it's pretty fast if you know how to set it up.
Paleo-Jow Forums was pretty smart in choosing what to hate on, because paleo-Jow Forums wasn't full of uneducated /v/edditors and btrfs was unstable back then, which was a damn good reason to not use it. Neo-Jow Forums is a hivemind that doesn't know shit about CS or IT and mindlessly touts decade old blogs as godspell.

>It's not stable enough for usage in data centres with thousands of PB of sensitive data
It is stable enough for your irreplaceable data, like original photos?

Yes. I use it for backup too.
Unless you change hw often and have very specific configurations, it is as safe as you can get without going for FSCQ in slow-mode.

I'm waiting for HAMMER2 to be ported to linux.

install dfly

Attached: ladydfly.png (1920x1080, 1.08M)

>says the guy using an OS with 1% marketshare

I might if I had a server. Besides, GPU drivers for 2400G aren't perfect even on Linux, even if there were no other reasons.

>BTRFS is good
>No, it's bad
How about you fags actually post reasons why you like or dislike it? I just installed fedora on my laptop using BTRFS with an encrypted / and would like some pros and cons.

>cmd-F
>APFS
>0 results
shame....

Somebody has to do the job since you neets obviously won't.

Is there even any difference for basic bitch desktop use?
Should I be formatting my SSD with yaffs or something?

Attached: 1498643635923.png (785x757, 375K)

Bank teller jobs?

Unless you are actually using some real form of RAID there is no reason to use anything other than ext4.

>Don't use stable things.
Why not just buy a Windows?

>It's neo-Jow Forums to care about stability and safety over blind progress
I have personally lost data on btrfs I have lost data on many filesystems. You know where I have never lost data unless the disk simply wouldn't mount? Ext3 and ext4. You say "it's totally stable now!". I'll give it 10 more years. There are other options if you need the snapshots and such. For a pc those are all meme features anyway. Ext4 is fine. It's boring. Sometimes boring is best.

>made by fagbook
Into the trash it goes.

>Sometimes boring is best.
You're a future OpenBSD user, user.

Yann Collet is great. He could work for the NSA for all I care.

>2018
>not using APFS

Attached: 1518923720565.png (179x282, 117K)

>and creating btfs-on-lvm is too much of abstraction and is going to be slow.
>uses one level of device mapper
>thinks adding another layer will make it slow.

you realize lvm is just an abstraction to raw device mapper autism right?

ťbh I've had discovered quite a few corrupted images on ext4. Though I do not know if the corruption is from before I migrated from ntfs-3g or after, or if it's just shitty HDD.

I've had several corrupted ext3 systems (bad luck on dead sectors) and two ext4 failures - both were due to laptop running out of juice and none was recoverable.
Meanwhile btrfs has survived all hw failures without data corruption and most without data loss.
Btrfs is safer than ext4 both in theory and practice, especially in case of hw failures. Even default config has stronger guarantees than ext4.

>an old Windows 7 install i use to burn live CDs

what
what about dd

I have only ever used ext4 on my SSD. Can someone explain to me what the benefits of other filesystems would be? I don't really know if I should bother.

Whomst here NTFS master race?

Dups confirm NTFS master race.

Attached: 1457962531888.png (480x480, 351K)

mfw I really enjoy working on my FreeBSD pfsense machine since it's so comfy, but I never need to because it's so goddamn stable.

Attached: 0e9.jpg (499x499, 24K)

And this is why you need checksums.

Guys i am an btrfs evangelist and use it actively since five years or so and it finally managed to crap out on me on a new partition with a freezing laptop. btrfs-fsck is a joke and should be put on the wall for how useless it is, I had to completely ditch that partition because it was unrecoverable.

I'm back to luks+lvm+ext4, ext4 may give me more broken files on freezes/power outage but at least it doesn't take all of my files into its grave.

Fuck btrfs, why is zfs still not merged yet?

Was Linus talking about you in China recently?

Were you the one sending bug reports to the kernel for a file system everyone thought was extinct??

Afaik zfs has licensing issues because it's free, but not gpl. Some distros support a zfs root install by just including the ported packages, notably proxmox.

Technically, it's not zfs that has license issues, it's linux with its gpl.

I agree with this. I use btrfs because it's a file system rather than just a journaling system and I've heard they fixed the bloat problems it's had in the past. But I would like actual reasons.
>inb4 it was xyz in 2002
That's not a argument faggot. We want to know it's current pros and cons. Getting wifi to work on linux in 2002 was a bitch, but it isn't that way anymore yet people still make that shitty argument.

Corporate sphere does not trust it much, and neither do I. It was written by people who didn't know exactly what they were doing, and with too much tyas. And while such approach will somewhat work if you have enough test subjects, more competently written FS will just be more reliable.

because they can't. btrfs is basically fud'd to shit by RedHat because RedHat employs XFS engineers, an even worse fs.

worst case, you have backups anyhow since no storage medium or filesystem is bulletproof.

The only real con i can see in the FS is the main sponser, Oracle, basically dropped it. it's still upstream though and it gets fixes and contributions.

hating something because of which company made it is no better than loving something because of which company made it

using btrfs as an example;
- can convert ext4 to btrfs in-place
- snapshot support (good for reversing 'accidents')
- compression support (save space, potentially improve performance, too)
- checksums (provides a gaurantee that what you read back is what you wrote)
- can be converted to and from raid at any time, even online
- subvolume support (special folders which can be mounted as if they're partitions, like LVM but easier. you can use them to say, keep multiple roots and a home on the same volume without needing to decide on how much space to give them ahead of time)
- can be migrated to a new disk while online
- no fsck needed during mount

-- oh, and importantly, a plain btrfs volume acts like a regular filesystem, you don't need to learn anything to begin using it, and all it's features can be turned on after filesystem creation, so you don't need to worry about missing anything

I wish the BSDs supported EXT3/4

Is there a way to find what data was on a drive before you turned on FS compression? I know I have a decent amount of uncompressed data on the drive and the only way I know of to fix it would be to move everything off and back on the drive, but that's 4TB of data

not sure about that, but you can tell btrfs to compress everything
btrfs fi de -rvczstd /
(assuming 'zstd' compression and '/' mountpoint)
this also defrags the files, since that's the main purpose of the tool. see "btrfs fi de --help" for other options

Would that work to switch from LZO to Zstd or do I need to run a different command for that?

yes, that can be used to change compression type
i've done that to 'upgrade' my btrfs volumes from lzo to zstd also

Perfect, thanks

np, remember to specify zstd as your mount option so new files use zstd

You have a issue with the way they approach developement and that's fair enough but what are the actual issues that it has? Is it more volatile with data (more prone to corruption, breaking, etc) than ext4 or zfs? Or does it just store things bad?

it's development was arguably rushed, and some things like raid5/6 support was greenlit while being critically flawed
people who tried it early on may have gotten burned by it, and it's tough to sell a filesystem when there's stories about data loss floating about
i admit i got burned by it back in 2012 when i first tried it, but i've started using it again and it's been fine (though i haven't had the guts to try raid5/6 again...)
filesystems are tough, and arguably the most important part of any system (since they handle the permanent storage of data, which is absolutely the most important part of any system), so of course, people want to be totally, 100% confident in them
anyone should have backups, regardless of their filesystem choice, but you know many don't anyway

-- oh, not the same person, btw

yes

What would you recommend instead? ext4?

no?

Attached: 1521785427975.png (500x505, 194K)

i'm using btrfs and mdadm, while keeping an eye on bcachefs

ps. and yes, i have used zfs
zfs is great at what it does, but it's very inflexible for a home user

-- to be clear, i haven't had problems with non-raid5/6 btrfs
and even without it's raid features, it still has many advantages over ext4. i don't think i'll ever use ext4 again

Ahhh ok. I see. What are you using to keep a eye on bcachefs?

the developers' patreon page gets regular updates regarding it's progress, it's been mentioned on phoronix a few times as well
i haven't tested it, since it's not frozen or mainlined yet, so it's not ready for any actual use, but it's proposed featureset is very appealing

-- also, being built from, and by the guy who made bcache, it has a decent chance of not falling on it's face
bcache isn't a filesystem, but it's the skeleton of one, which is one reason why the guy decided to flesh it out into a filesystem, and it's been around for ages, and it's very solid

F2fs is a poor choice for an SSD. It's designed for use on raw NAND flash and performs its own wear leveling. That will fuck with the wear leveling your SSD does, and not pass TRIM commands to it.

Give me one good reason I should use a meme filesystem for data I want to keep long term.
>ext4 is stagnant and doesn't get new features
So fucking what? Why do I want a filesystem that's updated all the time? That seems like the opposite of what I want, which is a stable, reliable, rock solid storage for my files.

this. "optimizing" for ssd's is harmful anyhow. literally everything has some FTL. I don't even know why linux still defaults to CFQ or deadline for io scheduling in 2018.

>reliable, rock solid storage for my files
ext4 is out then. Journaling doesn't do shit when hardware fails. You want checksums and CoW at the very least.