What will replace EXT4?

Butterfacefs

btrfs has already replaced it and it has quite a few very nice features. bcachefs might replace that some day.

reise

Attached: CopyQ.VMrpPX.png (397x274, 90K)

Hahaha

look it's candlej-

XFS is the future. Ironically it was also developed like 30 years ago.

>ZFS
>Need a gigabyte of RAM just to work
>XFS
>*fails after outage*
I hope the future has none of this shit, and by the popularity of these, I say we're on the right track.

Are you on ext2?

>Need a gigabyte of RAM just to work
new PCs are being sold with 8 gb as standard, 16 is becoming more common
phones are now being sold with 8 gb ram
sure you might not do well on a raspberry pi but for normal users, zfs is definitely doable

For some applications, maybe.

I feel like the SSD era is the wrong time to start developing filesystems with more CPU overhead, though. I'm sticking with EXT4 for (on average) better performance and relying on backups instead of fancy data integrity features, because you need backups anyway.

>Poorpad user

>have 30TB ext4 volume
>smallest possible block size is 1MB, literally wasting like a terabyte by now
>not allowed to use zfs because it's shit at random io
>butterfs still indev and too unstable

HAMMER2.
dragonflybsd.org/hammer/

Attached: 1545250460136.png (1440x1080, 1005K)

fpbp

this

>butterfs still indev and too unstable
imagine believing this

The cloud
Local file systems are obsolete

>quotas still fucked
>cannot into raid6, has no methods to close write holes
Yeah nah you're a cunt.

Ext3 filesystems also did not provide sufficient bits to store dates beyond January 18, 2038. Ext4 adds an additional two bits here, extending the Unix epoch another 408 years. If you're reading this in 2446 AD, you have hopefully already moved onto a better filesystem
opensource.com/article/18/4/ext4-filesystem

10 years later and btrfs is still slower and buggier than ext4
dead on arrival

Attached: 1560556157649.jpg (2433x3323, 1.47M)

>t
A lot of bugs were fixed. Most links anons post are backdated to problems as old as 2014.

Extf5

call me once they do. apart from glorified ftp servers there ain't much development going on in regards to file systems. as long as there are no security vulns there's no need to change anything.

Even if this was true, cloud servers still need a filesystem to run on.

its easy just dont use those features lmao

ext2 is a bit too simplistic and limited, but ext3 actually ought to be enough for anybody

>new PCs are being sold with 8 gb as standard
>8 gb is the "standard" since 2012
>7 fucking years the same amount of RAM is "standard"

I bet you'd prefer to keep your brain in someone's jar somewhere (local skulls are obsolete).

ThotPatrollerFS

>btfofs

fat8

2014 was like two years ago bro

8GB is still more than sufficient for 95% of normalnigger consumers.

I have encountered new computers being sold with 4GB of RAM still.

>I run my VMs on top of btrfs
Pls get back to us on how that worked out. This and the impossible to recover once full retardation are why btrfs is still ntfs tier

fuck off zoomer that's not how the meme goes. you have to say candlejack BEFORE yo-

stop meme'ing or being retarded

Lol that skyline looks like a 90s Cadillac coupe.

No matter who came up first with it, this is the most based automobile technology design

>XFS is the future
>can't shrink
>no features
>reliability issues

Only reason anyone talks about it is redhat is overloaded with so many xfs contributors.

It can though. Write holing only happens if you have a media failure after a power failure before a scrub operation. Where do you dipshits come from? All parity implementations are prone to write holing and anyone telling you otherwise has sufficiently low probability risk or outright lying to you.

Nothing, ever. At this rate Linux will die sooner than anything comes close to replacing ext4.
BTRFS is a steaming pile of shit.
ZFS has license autism and is out of tree so you can get fucked at any time, which means no enterprise use and less eyes on the code base.
XFS has uncertain future. RedHat decided to start bolting on BTRFS/ZFS features, so most likely XFS will become very buggy in the future.
JFS is on life support.
Reiser4 is as dead as Reiser's wife.
I'm starting to hope HP-UX will get open sourced at EOL time so we can get VxFS.

>XFS
not shrinkable

what will replace EXT5?

I'll keep using ZFS. Cause it just works and it works perfectly fine at keeping my data error free. Turn off Dedup to free up ram, speed things up a bit.

I've found that most cases of ZFS/Freenas failure is due to the builder (user) not building it right. You half ass something, it won't last long.

This is another complaint "I can't expand my Raid Z pool". Yes you can. What you can't do is go from a Raid Z1 to a Z2 config. Also If you took the time to see just what your capacity needs were you wouldn't have this problem. Which goes with the above "half ass" point I made above. You got 20TB of data now, you fucking create a 40TB or larger volume. End of story. There you've nipped your "growth" needs in the bud.

Drives are at fucking 14TB now. Sure they cost but if your looking for a small server footprint and you got a lot of data (or will have) guess what, time to whip out that credit card.

More ext4. Or possibly btrfs if they ever make it not lose data.

hopefully nothing, zoomers hate whatever has been around for too much

mount a btrfs folder as root in the vm. or disable cow and use big file as blockdev for the vm. both do well in performance.

EXT7

ZFS isn't a good idea to boot from, I dunno why the solaris fags over at freebsd got everyone doing it, it's more than just an array and therefore shouldn't be thought of as similar to other filesystems ontop of an array. And actually there's reflow tech going into ZFS that looks like it might do the raidz1->raidz3 migration everyone realizes is necessary when they go to add in their next vdev and finally see how little it takes to lose the pool now.
t. big data engineer who works with arrays on a daily basis.

bump for zfs

EXT5

kek

Except bcachefs will replace it and Linux is going nowhere as long as servers exist.

Attached: 1566431046539.gif (400x225, 975K)

anonymous is on point as fuck today

Attached: reiserfs.jpg (1076x662, 199K)

>>XFS
>>*fails after outage*
Fake news.

I wish they would let Hans continue coding in his jail cell.

Really? Is XFS a meme?
>tried to install linux yesterday on XFS
>error: can't install grub: unknown filesystem

>grub on XFS
Why? You use XFS only for the root and home partitions.

I use f2fs on all my SSDs. I'm not sure if it's beneficial on drives with encryption.

What will replace EXT7?

Inter Galactic Filesystem

Ext5 whenever it's released. Stop shilling the slow piece of shit that is Btrfs.

>turn off dedupe
>turn of COW
wow, fast

why wouldn't it be?

Whoever says btrfs is better than ext4 obviously never had a crash or some shit with btrfs. Also,
>phantom space usage due to corrupted metadata
>delete files but your used space is the same
Fuck btrfs. I wished i hadnt used it for work. Fuck btrfs for real.

>refering to bugs from 2014 that are long since fixed

EXT6

>ZFS
rock solid, okaish perf if you set it up right. great for a few use cases, pretty shit as a general file system.
>XFS
More minimal than ext4, but redhat is working on fucking it up. really good as a general filesystem
>HAMMER2
better than zfs [citation needed], only available on a select few flavors of bsd
>NILFS
God of low latency low throughput
>reiserfs
Good at low latency high throughput operations on small files.
>REISER4
rieserfs killer. The moment it becomes in tree no one will talk about reiserfs anymore.
>f2fs
fast file system for flash storage.

You use the right tool for the right job.
/ - NILFS
/var - REISER4
/home - f2fs (if i was using spinning rust i might use xfs here)
my NAS is using ZFS, as well as my vm server where fs perf doesn't really matter.
I want to try out HAMMER2 at some point though.

/ = ext4
/home =ext4

totally valid for your average signing rust hard drive, perhaps your most intelligent choice

At that point just
/ - ext4
Is perfectly valid
Maybe also
/ - ext4
/boot - ext2
I partition and format according to use case though and this scenario would likely never come up since all of my client machines have an ssd.
A few machines have a second larger spinning rust drive, which depending on my mood/use case might get formatted ext4, xfs, or UDF.
For 99% of servers the default is fine, especially if you're vitrualizing on top of ZFS or something similar like I am.

TempleFS

Yes. The right answer, most people don’t like the fact that it’s that simple.

Also,
UDF? Why tho, just curious.

Dual booting
It's cross platform and has no fuckery with permissions. Pretty much every OS that has DVD support can read and write to a UDF drive.
I use it on all of my flashdrives too
I have 1 system I use for lan's to play old shit like UT99 and jedi outcast every so often, and debug windows client issues. I'm dual booting until I can get the extra time/money to configure pci passthrough.

ZFS *for what it does*, if redundancy and protection against silent data corruption are important to you. Maybe Btrfs if they iron out the problems they still have with it.
If you're obsessed with performance and low overhead then something else probably.

Ummm... I have bad news for you if you depend on XFS. I lost terabytes of video footage kept that shit after a power outage. The data was there, but scrambled. Every damn video had some corruption.

I like xfs, so I hope this is eventually implemented.
xfs.org/index.php/Shrinking_Support

ocasional reminder that ext4 IS journaling done right. There was no need to go for xfs.
If you value your data, stick with ext4

Just use mdadm and not btrfs raid