it's gay

It's 1990s technology and subpar like all of windows inside and out

>>109409651
It #JustWorks
Other filesystems overcomplicate things to the Nth degree.

I like that for some reason Everything can use it to search the mft very fast. For some reason no filesystem on linux can do that.

>>109409651
It's inferior to Btrfs. ReFS is a meme that literally breaks between windows versions, lol.

>>109409680
You can do it with any journaling filesystem, but only a retard would expose it to non admin users.

>>109409670
Pretty much this. Windows is good at one thing: integrated solutions. The actual "how" is all dogshit though.

>>109409651
>xfs
What's the deal with xfs anyway? It's been around forever, but does anyone actually use it over ext4? And if so, why? What is even the point of xfs?

>>109409651
It's niggerlicious.

>>109409752
It's less shit and also has copy-on-write at the file granularity level, which means working reflinks. Only shit deal is you can't shrink an xfs partition. Ext* has always been a joke fs with a lot of weird hacks and sloppy programming. Sadly vfs is infected by it and thus everyone suffers.

>>109409694
So it can't, got it.

>>109410057
fuck you

>>109409651
>How does /g/ feel about the ntfs filesystem?
Very robust, a lot of features.
>How does it compare with ext4 and xfs in your opinion?
Weaker file systems with the benefit of it being faster.

Doesn't matter if your drive dies your filesystem is meaningless, it can have layers of backups and protection but all of that becomes pointless

Also loonix is shit

>>109410184
moron.

>>109410219
Dimwit

NTFS is dogshit, just like everything else about Windows. It's long overdue for a replacement but Microshaft is too stupid to create anything else that works. And no, "AI" won't save them.

>>109410275
Ai will save loonix though

>>109409651
You need to defrag NTFS while you don't for ext4

Never had a problem with it on Windows but linux support for ntfs is still dogshit
>use kernel module for ntfs
>corruption up the ass
>use fuseblk/ntfs-3g
>No issues but have to wait 9 years to write a large amount of data
I should really get another external 2 TB drive so I can temporarily move my shit there and reformat my current external as exfat or something but fuck these prices.

>>109410320
>Using a HDD in current year

File: 1754864860752480.png (450 KB, 1761x1250)

450 KB PNG

>>109409680
what about this?
https://github.com/cboxdoerfer/fsearch

>>109410320
Defragging gives me time to meditate

>>109409651
NTFS is actually a much better filesystem than it often gets credit for. It has a reputation among Linux users as "that Windows filesystem," but that's mostly because it's proprietary and not because it's technically poor.

All three are journaling filesystems, so they all recover well after crashes.

NTFS:
Metadata journaling
Mature
Proven over decades
Windows automatically runs recovery when needed
On Windows:
Performance is generally comparable to ext4 for everyday workloads.

On Linux:
The modern ntfs3 driver is much faster than the older FUSE-based ntfs-3g, but it's still not as well integrated with the Linux kernel as ext4 or XFS.

ext4:
Metadata journaling
Extremely mature
One of the most reliable Linux filesystems
Great for lots of small files
Good for databases
Good for VMs
Good sequential performance
It's basically the "Swiss Army knife" filesystem.

XFS:
Metadata journaling
Enterprise-grade
Designed for very large filesystems

Optimized for:
huge files
streaming
media
parallel workloads

are where XFS shines. It also scales to enormous filesystems.

I'd trust all three with important data.

Linux permissions s is one of NTFS's biggest weaknesses under Linux. Linux has: UID, GID, chmod, chown, POSIX ACLs. NTFS has Windows ACLs. Linux has to translate between the two. That translation works... mostly. But it isn't perfect. You'll eventually encounter weird permission behavior if you're using Docker, Samba, or NFS.

>>109410411
Windows historically fragmented files more because of how space was allocated. NTFS therefore includes a built-in defragmenter.
ext4 and XFS generally fragment much less. XFS in particular does an excellent job with large sequential writes.

For disk repair, ext4 has fsck.ext4.and xfs has xfs_repair, which are both excellent. NTFS has chkdsk, which has improved a lot over the years, but it can still take a long time on very large volumes.

NTFS supports transparent per-file compression. It's easy to use but uses relatively old algorithms and isn't especially efficient. ext4 has no built-in transparent compression. XFS also doesn't provide transparent compression in its standard implementation. This is one area where filesystems like ZFS and Btrfs stand out, offering modern compression algorithms such as Zstandard that are often both faster and more effective.

>>109410344
This is pretty nice, gonna use it from now on. Still, it took it quite a few minutes to index all my drives on first run whereas Everything on Windows was some black magic where even on first boot of the software it took only 10 sec to go through my shit and then anything I typed in would come up instantly.

>>109410612
it is still a mistery to me why MS hasn't bought/bribed/integrated Everything into windows yet, as imho it is one of the most basic and most useful windows apps out there
similarly notepad++ or paint.net

Why is apple filesystem so much better than ntfs? They are both proprietary, yet apple is so much better and more performant.

>>109409680
Linux tool 'locate' is rather efficient but has to have an updated database.

>>109410674
iCucks don't even know what a filesystem is

Writing to ntfs from Linux, especially if you’re using an external drive formatted as ntfs, is peak russian roulette
I do like instant seeking and that data recovery tools are pretty good. But it has nothing for detecting bitrot, much less self healing. Could be worse, could be better.

>>109409651
It works, but is hopelessly outdated just like ext4 or xfs.Can't imagine using a filesystem that doesn't do data checksumming and CoW like Btrfs or ZFS

>>109409651
It's unsuitable for any serious usage. Despite NTFS5 being journaled it shits itself every few years without fail.

>>109409673
NTFS was the most complicated FS ever made by a huge margin until ZFS finally overtook it 12 years later. Which was basically a lifetime the way tech moves these days. It was and still is ridiculously overwrought for what most Windows users do with it.

>>109409780
What file system does templeos use, genuinely curious

>>109412702
>>109412748
Try not buying SSDs from samshit and aliexpress.

>>109409694
it's done using the MFT, not the journal

>>109415093
If you're clueless about the technology, ask a question and stop posting nonsense

>>109415119
Sorry can't hear you over all the ntfs errors I'm not having.

>>109410459
modern (win10+) ntfs compression (xpress) is decent but it's not exposed to the gui. the gui option is for old ntfs compression which is hot garbage
only btrfs supports zstandard

>>109412748
ntfs doesn't support data journalling

>>109410057
>>109409694
>only a retard would expose it to non admin users
Linux and its obsession with "security" is holding everything back. Imagine trying to enforce an arbitrary permission boundary between Muh User and Muh Kernel when there's only ONE FUCKING USER on the computer. Linux devs haven't managed to drill into their stupid heads that most of the world no longer uses 1960s style Unix mainframes where multi-user was a real thing.
As far as I'm concerned, the only reason for a user-kernel separation these days is so that pajeet apps that crash due to bad code don't crash the whole OS.

>>109415147
Sometimes pajeet apps are co-opted by malicious remote actors if not pajeet himself.

>>109415124
cool thing about ntfs is chkdsk will say everything's cool but it can't check for data errors at all.
if you've ever fixed a steam game by checking it's files in steam... i have bad news for you

>>109409651
I use btrfs.

>>109415162
Once like 10 years ago, and it was in response to a known error. Regular file corruption isn't normal if you have working hardware. I ran solamente btrfs on my laptop for like 5 years and realized I was being retarded because there were zero checksum errors.

>>109409651
It's excellent. Better than EXT4 because pretty much every system can read or write to it and it has built in features like transparent file compression.
I think Btrfs is the best file system Linux has.

>>109415158
So? How does allowing a malicious app to read the FS journal or MFT have any impact whatsoever on security? They can just read all my files and passwords directly. There is simply no way to guard against running malicious code on the local machine. Things like Spectre and Meltdown have proven that even if you have perfect software you can't do it.
The solution is to stop pretending that the OS or muh sandbox or whatever will protect you against running viruses on your computer. Just don't run viruses on your computer. If you think you can't manage that, then don't use a computer. It's literally that simple.
Meanwhile, if software can read the MFT or journal or whatever it reads, it lets me search my files instantly. If I disable the Spectre and Meltdown bullshit, I get more performance. So obviously that's what I'm going to do. Not gonna cripple my PC for some imaginary security.

>>109415186
>ntfs is better cos linux can use it
it's ok but not suited for long term usage since there's no chkdsk for it
also there's a btrfs driver for windows, too

>>109409651
I feel like it's better to decompose the acronym to NT File System. From what I recall, NT is removed from its original initialism these days in the Windows sphere. It's a bit like if you called it "GNU's Not Unix/Linux".

>>109415135
>only btrfs supports zstandard
zfs supports it

>>109416006
oh, it didn't last i checked

File: 1451952058855.jpg (92 KB, 500x617)

92 KB JPG

>>109415186
>and it has built in features like transparent file compression

>>109416049
Are you retarded?
It does. And with modern compression algorithms, too, like XPRESS8K or LZX. Try that in EXT4. What are you gonna do? Mount a FUSE FS that reads from a .zip for maximum ass-tier performance and usability?

>>109416643
So why isn't it exposed in the GUI, cuck? Why does Windows still pack the old outdated garbage in the disk GUI?
Also no zstandard. ntfs compression is overwhelmingly inferior to what btrfs and zfs offer.

>>109409651

>>109416643
NTFS compression is fucking useless. when I want a compressed filesystem, I take squashfs with any of: gzip, lzo, lz4, xz, zstd.

>>109410344
>gtk

Apple File System is basically all the good things from both NTFS and ext4.

>>109415210
>bro just run dos
Ok thanks for the dissertation on your own incompetence.

>>109409680
MFT fragmentation made it much worse than plain FAT32.

>>109409651
works great, never had a single problem with it (other than a silly path length issue from a very long filename in too many subfolders)
even with hardware issues of the drive data recovers pretty well from it

>>109418658
Apple File System is copy on write. Why are you even comparing it to journaling filesystems? The comparison is with zfs and btrfs.

>>109409651
Garbage performance. Extremely quirky. But overall good.

Was ReiserFS good back in the day?

>>109420501
There was a period in the early 00s it was the go-to if you knew what you were doing. Kind of like xfs is now.

>>109409651
slow and bogged down with random super niche features that nobody asked for

>>109420750
Fucking explain yourself

>>109418237
Just use KFind.

>>109409651
I’d like it if it was more resilient to mft corruption. Which it isn’t. Lord help you if you ever think ntfs is an acceptable compromise for an external hdd that’s supposed to work on windows, linux and macos at various points because that sucker will get corrupted before long.

If you’re a wincuck then ntfs is probably fine

wtf is an “ACL” anyway? FAT or EXT3/EXT4 are fine as long as I can have my Paragon drivers on macOS

>>109425017
They're like slightly more complex Unix permissions where you can set access for specific users/groups other than the owners. Plus delete permissions and more complex inheritance rules. Pretty much the only use case for them is making group shared directories where users won't accidentally make things other members of the group can't modify or delete. Because ACL masks set the correct default group ownership and {f,d}mask.

>>109409651
It's pretty much just ext4. By default it has journaling, but only for the most basic of file system info so that a sudden crash or power outage doesn't completely brick your install. It's very basic and thus fast, but it's not useful for serious data storage.

>>109409673
>is mogged by simpler and more complex filesystems
NTFS is probably the most convoluted piece of shit I ever did see

File: 16576-acl-tear.jpg (92 KB, 800x936)

92 KB JPG

>>109425017
like xattrs except think selinux tags

>>109410411
None of these filesystems are good enough for big boy data because none of them are CoW and none of them can heal. Ext4/NTFS/xfs have no idea about your files past some very basic info. If you lose a drive with any of them on it, you basically automatically have to turn to recovery software and/or third party solutions. Real filesystems with proper redundancy and healing (ZFS/Ceph/BTRFS/etc) can do proper RAID that self heals and takes instant perfect snapshots because the filesystem knows everything about every file on it down to every byte and keeps track of them. Losing a drive in a RAID config with any of these systems usually means you can just take it out and shove a working one in and all your data will be fine.

>>109425675
nice try sneaking btrfs into that list of "real filesystems" kek

>>109425679
Ok sure, BTRFS is not on the level of ZFS or Ceph, you got me. I threw it in there because it has many of their useful features which cannot be denied. The fact that is still technically in beta does still, unfortunately, make it more of a joke fs.

bocchi the rock file system is good and I’m tired of pretending otherwise

How’s btrfs on winblows? I want to support both linux and win but dont need full zfs autism

>>109409651
Slow for gaming compared to EXT4..

Ntfs is a superior filesystem.

>>109410131
ext4 is probably not faster. People mistakenly attribute perceived performance to the file system whilst it mostly comes from the drivers and OS design. IO on windows is notoriously slow regardless of file system.

>>109410320
This statement is just nonsense. Its fundamentally impossible to create a file system that never needs defrag.

>>109426218
ReiserFS says hello

>>109409651
its absolutely atrocious, the worst piece of shit ever created. ntfs fragmentation is the sole reason computer shops even stay in business. the performance is so dogshit, compiling code is x3 faster on linux, unironically. I tested this with the sqlite codebase, 3+ minutes vs 54 secs (windows vs linux)

>>109409690
Is btrfs dying, though? It's not included with Red Hat. I was surprised to see that AlmaLinux started including btrfs, but I had already set-up ZFS.

As far as I know, btrfs is included with Fedora and Suse. That's it.

>>109426527
> compiling code is x3 faster on linux
This has basically zero to do with NTFS. Windows itself is the problem. Windows has massive overheads for IO operations and process creation.

>>109409651
Not The File System for me

>>109426670
>>109426527
Way way way too many variables involved to point a finger at NTFS there chief.
Same compiler? (No)
Same antivirus? (No)
Same operating system (No)
Same #includes? (No)

Come back with a proper test next time.

>>109427342
Why did you reply to my post?

>>109425675
In a lot of cases integrity is either the hardware storage controller's problem or largely handled by the application.

>>109428196
But why should anyone use a journaling filesystem over a cow filesystem in this day and age? what is the usecase for journaling?

>>109428392
Not having outsized CPU overhead or uncontrollable fragmentation.

>>109409651
another 'microsoft version' that is worse and became standard

>>109428472
The fragmentation is indeed an issue, but it only really affects speed, and it hardly affects speed if you've got filesystems on SSDs. For data security CoW is superior, even if it has warts.

>>109432116
Yeah but it's really easy for SSDs to do IO faster than the CPU can do integrity. Simple data only CoW like XFS is good though.

>>109409651
>lets fragment data so it reads faster in the HDD
>everything becomes slow

>>109425689
a joke filesystem that facebook uses extensively? only raid5/6 isn't complete because they don't care about it

>>109432821
Even btrfs raid5 is good enough for home NAS if you understand the problems with it. zfs users are 95% larping the mid-late oughties systems analyst they never go to be. Real enterprise storage went back to hardware acceleration and dumb filesystems when SSDs killed nearline HDDs.

>>109433465
i don't disagree, as long as you follow the recommendations that are use raid1 for metadata (raid1c3 for raid6) and scrub immediately after an unclean dismount, you should be fine

>>109432821
and it’s in the kernel, that’s kinda a big deal for why you would use it over zfs

>>109433560
only thing i wish it had that zfs has is l2arc. i've done btrfs over bcache before, but these days i have enough ssd storage to just put everything besides media on ssd's entirely

>>109409651
>1 TB HDD, NTFS
>run it at around 99% full for a year, removing and adding files
>horribly fragmented to the point where it can't even play 1080p video stored on it without stuttering
>defrag moves some stuff around and leaves it about as bad as it was
>try some 3rd party defrags, they either do about the same or say it'll take about a week
>give up, copy files off, format, copy files back, finally it works well again

>2 TB HDD, ext4
>do the same thing with it
>6 years later the HDD is dying, get a new HDD and dd the filesystem over to it, preserving any fragmentation
>another 4 years later, it's still working just fine
>check fragmentation level out of curiosity
>under 2%

So yeah, I like ext4 better.

>>109433580
Why the fuck don't you store backups on HDDs?

>>109433899
>1080p video
You should be using xfs.

>>109435094
I'm not even talking about dealing with raw footage or scratch for editing. I'm just trying to play normal mp4 files, but my disk was so fragmented it couldn't keep up. And it's NTFS because it's on a Windows machine.

>>109435083
i'm only talking about hot storage, i didn't think that'd need to be specified

>>109409651
They are all good at storing files.
Why should I care more than that?