EXT4!

What a fucken joke is EXT4?!

File system limitations
EXT4
Max File
16TiB (HARD LIMIT)
Max Volume
50TiB (Soft limit)
access.redhat.com/solutions/1532

So if you buy some 16TB (8x16TB = 128TB) HDDs and put them in a raid you are reaching the limits of EXT4.
The real insult is the Max File limit.
Because I use insanely large files so one file is expected to fill up 99% of the HDD.

Meanwhile on windows (win8)
NTFS
Max File
256TB
Max Volume
256TB

In comparison win7 did have a volume limit of 16TB
WHAT THE FUCK!

Why is there no EXT5 or whatever, give me a increase in the max file limit!!!!!!

Attached: 2.png (915x488, 140K)

Other urls found in this thread:

access.redhat.com/solutions/1532
newegg.com/Product/ProductList.aspx?Submit=ENE&IsNodeId=1&N=100167523 600376738
xfs.org/index.php/XFS_FAQ#Q:_Why_do_I_see_binary_NULLS_in_some_files_after_recovery_when_I_unplugged_the_power.3F
en.wikipedia.org/wiki/Ext4
wiki.archlinux.org/index.php/file_systems
en.wikipedia.org/wiki/Copy-on-write
twitter.com/SFWRedditVideos

>ext4 is the only filesystem that exists
lol wat?

>The real insult is the Max File limit.
llvm a shit

Are you dim? The maximum filesystem size for ext4 is 1EB

God never intended for you to have more than a few gibibytes.

Suggest a better one for loonix.
EXT4 cancer is everywhere, and its the default.

How large can you get a file in EXT4?
>The maximum filesystem
See actual documentation
access.redhat.com/solutions/1532
And
>Max Volume
>50TiB (Soft limit)
>(Soft limit)

XFS
>max file size
8EiB
>max volume size
8EiB

If you need a bigger limit on your hard drive go format it into something else

>ext
>ntfs
What fucking year is it

I was thinking about it
1) Is that thing stable?
I don't want to get massive data corruption.
This is not a grandmas facebooking computer, where the OS and everything can corrupt itself and you only reinstall.
My data is extremely important.

2) Not the default on many tools.
Example i need to install this sewperatly and Vera Crypt tells me I can go fuck myself because it will literally not create a XFS volume only
EXT 2
EXT 3
EXT 4
FAT

What FS do you use?

>stable
It's been around for decades for use on big iron, it's stable
>defaults
I don't use anything FS-specific so I can't help you there

How many 16+ TB files do you have?
What kind of system produces data that large in a way it cannot be separated into smaller files?

its more a rant on Veracrypt and it blindly only using these file systems.
>How many 16+ TB files do you hav
Normal strategy
1)Buy a HDD.
2) Use 99% of it to create a file container where you keep your encrypted file system.
The benefits to this are no one will accidentally format your encrypted data.

>16+ TB
its 2018 you can buy 12TB HDDs get 2 of them(combine into one RAID etc) and now EXT4 can not have over 16TiB file container you created.
newegg.com/Product/ProductList.aspx?Submit=ENE&IsNodeId=1&N=100167523 600376738

>It's been around for decades for use on big iron, it's stable
I looked into it and in some forums people say it can zero out your files.

Imagine being this stupid

Imagine being so retarded you literally contradict official documentation.

xfs.org/index.php/XFS_FAQ#Q:_Why_do_I_see_binary_NULLS_in_some_files_after_recovery_when_I_unplugged_the_power.3F

access.redhat.com/solutions/1532
Show me where i'm wrong.
I'm waiting.
Do you evne understand half of the words used there?

Why not use full disk encryption instead of formatting it in ext4 and putting an encrypted file on it?
Seems dumb.

>Since XFS does not write data out immediately unless you tell it to with fsync, an O_SYNC or O_DIRECT open
>unless you tell it to with
literally joke FS.
Useless on anything other then a facebook computer or a server farm where you have redundant power supplies.
I can not gamble with my data and hope I never get a power interruption.

Thanks for the link.
it confirmed my suspicions about XFS.

>Why is there no EXT5 or whatever
ext4 was intended as a stopgap till something better came along.
>give me a increase in the max file limit!!!!!!
Nobody's stopping you from using a different FS.

What's wrong with adding a few characters to your fstab?

>Why not use full disk encryption
I was thinking about it.
It still is a think I need to make because of the HARD FS limitations not as a choice.

>Seems dumb
Benefits to container

1) no accidental overwriting of a important data.
What was on this HDD, wait was it important?
I can use first FS to simply write a note what the data is(create TXT) or literally
>DELETE_THIS_HDD.txt

2) it ensures that if my filenames are to long I don't get
>Can not copy file path to long
error.
I remember them from windows and I simply got into the habit because it fixed this problem.

> stopgap till something better came along.
The world is still waiting
>using a different FS.
See thread.
Name one example
XFS is confirmed to be a joke FS by its own developers.

ZFS, Btrfs

en.wikipedia.org/wiki/Ext4
Limits
Max. volume size

1 EiB

KYS OP

>ZFS, Btrfs
Using a COW FS where one file takes up 99% of the HDD.
Now that a terrible idea.

How retarded can you people get?
>wikipedia
VS
access.redhat.com/solutions/1532

>Max. volume size
>volume
Homework learn the deference between
Max File
and
Max Volume

And demonstrate to me how EXT4 can get files over 16TiB. Files not Volumes.

Attached: 1520574079802.png (657x539, 110K)

>Benefits to container
You seem to be using it wrong then.
A container is meant to be an encapsulation of a program so it doesn't touch the rest of the system.
Giving it a full harddrive array as a single file is pretty risky and there are better practices.

1) the limit is 1EB
2) if you use FILES more than 16TB you are doing something wrong..
There is no FUCKING reason not to split your output at 1TB for example (except being a retard)
Whatever you are doing that needs so big files is bad design/designed to fail so you might as well kys

We literally use it for all our 100+ bare metal machines for Elasticsearch. It would be known it wasn't stable long ago if that were the case.

>if you use FILES more than 16TB you are doing something wrong
Literally on the same level like
>640K is more memory than anyone will ever need

Or simply look at FAT32 the 2GiB file limit is hard if video files get over 2GiB today.

I work in this way.
Also have fun not having the ability to create VMs with HDDs over 16TiB. Because the VM-HDD is one big a file on the HDD.

Also COW FS are for normies not for VMs how exactly will it work if you use COW on 1 big file that constantly updates?

A) ultra lag and VM is unusable
B)
>I'm not writing this to the HDD until you give the command LEL.
And when the power is cut your VM will be in some ancient state after starting the computer.

Attached: 1518700081051.png (488x463, 28K)

Why are open source developers so retarded that they are doing everything they can to get people onto Windows?

Attached: 1518760563996.jpg (988x1059, 160K)

>1) the limit is 1EB
Literally doubles down on his retardation and not understanding the difference between MAX FILE and max volume.
The discussion was mostly about MAX FILE limits.

>>if you use FILES more than 16TB you are doing something wrong
>Literally on the same level like
K is more memory than anyone will ever need
He is absolutely correct. If you're using 16TB in size (or more) files, you're doing something completely wrong on all levels, there's no justification. Go back to the drawing board because it is a terrible idea even in theory.

>Also have fun not having the ability to create VMs with HDDs over 16TiB.
Why would you be attaching large volumes to VMs? Use networked storage. If your VM is doing something intensive, needs to have as little barrier as possible and needs those resources to begin with (and more) it shouldn't be a VM at all, it should be on bare metal.

>And when the power is cut your VM will be in some ancient state after starting the computer.
Get a fucking $200-300 UPS, problem solved. You're the one being cheap. These aren't issues, if this is shit you actually care about, you'd at least have a cheap, but reliable UPS anyway.

Are you just brain dead?

>you're doing something completely wrong on all levels
Literally making excuses for the retarded limitations of EXT4.

> Go back to the drawing board because it is a terrible idea even in theory.
You literally can not use a VM with a VM-HDD over 16TiB. In the next years this will be exactly like the 2GiB limit in FAT32.

You realize EXT4 has officially reached its limit at this point, right?

In the next 10 years 1TiB video files will be the norm and the 16TiB limit will look like a joke or FAT32.
Can you imagine having a 1PiB HDD that can only save 16TiB files?
Or a 1EiB HDD with 16TiB files?

> terrible idea
Want to make a calculation the size of a 2h (HD or 8K) video file that is not using lossy compression?

>Get a fucking $200-300 UPS
>Buy more hardware because our FS is retarded and unusable.
>Also we are not making EXT5
Funny how MS can do all these things with NTFS while you are only producing excuses
>You don't need files over 2MiB
>Buy more hardware because COW FS are literally pointless and give no security from crashes whatsoever

The real question is why are you defending this retarded situation instead of demanding a real replacement for EXT4

Attached: 1_yEcccYW4yRWUp_4UK_WGmg.png (580x282, 289K)

Not him but I use btrfs, I really like the snapshotting feature. Though it's definitely not the most performant fs and not for universal use, but for root I really like it. /home is XFS.

>1) Is that thing stable?
It was the default f/sys in rhel6, so that says a lot, especially choosing it over btrfs.

>XFS
The developers themselves say it will shit all over the HDD if you crash the system.

Haven't heard about that. Seems really good and performant to me. I guess for the most solid fs you'd probably be best off with ext4 but XFS seems to be more popular as an alternative by the day.

>btrfs
Terrible at working with large files (See having any VMs).
Are there other gotchas?

Like needing special commands or shutdowns for any of the data you edited being saved?
Because its pointless if the FS will lose all your recent work.
Reminds me of the
>My computer crashed and I lost the files i was working on
Slogans from the 90s.

If you are using delayed writes you can equally not have a FS or use something as corruptible like FAT32 Because every idiot can use a read only FS and then flush things to it.
The point of the FS is to prevent data loss if the system crashed (power loss, system panic).

I did rumors and then some user ITT decided to link to the devs and it turns out its a openly stated fact.

xfs.org/index.php/XFS_FAQ#Q:_Why_do_I_see_binary_NULLS_in_some_files_after_recovery_when_I_unplugged_the_power.3F
>Since XFS does not write data out immediately unless you tell it to with fsync, an O_SYNC or O_DIRECT open
>unless you tell it to with

Hope you like finding out all your files are full of zeros and nothing was saved from your session if the system crashes.
The idea of a FS is to give some safety in the system crashes every FS can say its perfect if you never crash the system.

>VeraCrypt
Dude, if you're storing files that are mission critical on arrays like this you shouldn't use fucking VeraCrypt. My personal advice would be to use LUKS to encrypt the entire disks and then create partitions in whatever format inside the container. You can also create raid arrays that way although I've never done that so no idea how that works.
But yeah, ext4 obviously isn't the way to go with this. Use ZFS or XFS or find something that suits your needs here: wiki.archlinux.org/index.php/file_systems

>Update: This issue has been addressed with a CVS fix on the 29th March 2007 and merged into mainline on 8th May 2007 for 2.6.22-rc1.
???

> XFS seems to be more popular as an alternative by the day.
I think its because of the hard limitations on the max file size.
Depending on who we are talking about...

You can not seriously say
>Database file over 16TiB? You are doing everything wrong! You don't need a 20TiB file!
And be taken seriously in server farms.

Also I think servers can afford to have redundant generators so the system will never be crashed from not having electricity.

>Suggest a better one for loonix.
XFS/BTRFS/f2fs

I'm sure there are some other gotchas too. I just use it for the snapshotting feature. I have never had my data corrupted etc in my life (outside of USB sticks I think) but I've had my system in a broken state, so snapshots are obviously going to be a priority for me.

>LUKS
Not audited, unknown, lack of basic features (why even encrypt your data if you are using LUKS ?).
VeraCrypt
audited proven tested, advanced functionality(like using files to encrypt volumes).

Limitations exist for a reason. To prevent retards from doing something they shouldn't.
>1TiB video files will be the norm
False.
If anything people started having less storage on their computers nowadays. So for normal use (99% of people) this is irrelevant especially when moving to streaming chunks rather than storing files.
If you're in a business using oversized files then you'd have a solution for it already, proprietary or not. Seeing how you're just a hobbyist and one of less than 10 people who find this to be a problem your opinions are irrelevant. Just use a different fs. EXT4 isn't made for this. You're saying a spoon is shit just because it can't be used as a screwdriver. Stop making dumb complaints.

Why do they keep the old description of:
>XFS journals metadata updates, not data updates
in the FAQ?

Looks like its exactly how its supposed to work.

>XFS
>ZFS
>btrfs

People interested in data center file systems wrote new ones instead of extending EXT further.

EXT4 is for consumer devices.

>False.
>500TB no one will use 1TB video files.

Sounds retarded to me.
Also I don't think you understand how shit today video formats look after they butcher the video with the compression.

The difference will be like watching a movie in GIF format.
>streaming
you realize if studios simply embrace the 1TB+ files steaming will look like watching GIF format video on the internet (remember thous days) , right?

>Limitations exist for a reason. To prevent retards from doing something they shouldn't.
This has to be the most retarded thing ever written.

Do you have files over 2GiB?
Well if yes then you must be retarded! And doing it wrong!
Because when Ms made FAT32 it created the limitations for a reason. To prevent retards from doing something they shouldn't.

You want me to list even older FSs to make you look even more retarded?

Attached: 1524411089805.png (960x730, 200K)

you can mount NTFS in *nix you degenerate

>NTFS
Literally admitting that all Posix file systems are shit and mounting the MS NTFS file system.
Literally admitting everyone in this *nix community is incompetent.

is this the FS equivalent of using wine to run windows software? Only worse?

>degenerate
You realize what abomination it is to use NTFS on Posix?!

Here's your (You). (You) really don't deserve it, but (You) seemed desperate.

If you have to argue about the pettiest shit regarding data preservation, your data is much less important than you want others to believe. Hard drives tend to get unreadable blocks or fail outright, and, depending on encryption mode used, a bad block might fuck up a few more blocks in your volume. Learn proper data preservation practices and/or stop pretending.

Not even a answer.

To sum it up:
>you are retarded
I prove the limitations exist
>no one will ever use files over 16TiB
I demonstrate this is wrong and only excusing EXT4
NTFS can manage this while EXT4 fails
>use NTFS
Literally admitting everyone in this *nix community is incompetent.
>Here's your (You). (You) really don't deserve it, but (You) seemed desperate.
Incoherent meme spamming shitpost.

Well but from the looks of it you don't want to use anything other than NTFS.
You could go ahead and set up ZFS, JFS, ReiserFS right NOW but you aren't doing it, so I'm suspecting you to just be a filesystem shill

#define MAXPATH 1024

>Well but from the looks of it you don't want to use anything other than NTFS.
Why is there no EXT5 with larger file capacities? Or something like this?
NTFS is capable of this so its not that journaling file systems hit some technological limit.

>ZFS
COW unusable for large files or has gotchas.
>ReiserFS
Are you serious?
Practically abandoned FS and some of the documentation says its limited to 8TiB volumes MAX.
I have no idea what is going on here mainly because ReiserFS development is a ghost town.

>You could go ahead and set up
look I'm not some autist who installs a FS and then waves his hand after the FS fucks up deletes files or corrupts them because he uses the computer only to shitpost.
I need to ensure some level of data protection, and this excludes all FSs who have gotchas where nothing will be saved if the system crashes.
FS who only save on proper system exits or upon giving special commands to the OS are useless.
Real people in the real world need to have the certainty the data I'm working on will not be lost.

>Why is there no EXT5 with larger file capacities
Because it's not a real problem that anybody is having.
Because there is no point, when other modern filesystems already exist.

What are you even talking about?

while we are bitch about ext4, the way it was explained to me is that ext4 never needs defragged because when you save a file, it saves the entire thing together, basically defragging itself on ever save.
So the drive is constantly doing unneeded writes.
this kills the ssd.

>why is there no consumer filesystem for storage server-sized files
just use ZFS or get your file chaos under control

>he doesn't know
The limitation of the Windows filesystem API is that no file path shall be more then 1024 characters. Unix does not have this problem.
Unix has a different problem though, it's that file names can't be longer than 256 characters. So you can't store all those booru tags of your fap pics in the filesystem. Not that you could on either system.

>other modern filesystems already exist.
You are talking about NTFS?
Answer me this are you admitting MS is superior and created a superior FS in 1993 ?!
NTFS was created in 1993 while EXT4 was created in 2008 and is more shit then NTFS?

Are you actually saying this?

I want this clarified.

>or get your file chaos under control
you don't even understand the difference between number of files and file size.

This thread is a mess

Since when was the quality of a filesystem determined by its maximum file size?
If I made a filesystem with an unlimited file size (up the the available storage space, of course), would it be infinitely good?

you're trying to drive a big car into a too small garage. if that aint chaos

Totally unrelated poster. Can you give the noobs a hint? I had synced all my important files via syncthing across multiple devices and just copied and deleted everything else to external hdds. Riddle me fiddled, I don't know jack about compooters.

>Since when was the quality of a filesystem determined by its maximum file size?
When in the near future its limitations will be crippling and in the next year require a complete rework of the model I'm currently in.

You can not seriously think that a FS who would not allow for files larger then 16KiB would be even acceptable today. Because this limit would be a pain to use.

If you are actually hitting the limits of ext4 you probably shouldn't be using ext4 anyway I don't understand the problem here.

>If I made a filesystem with an unlimited file size (up the the available storage space, of course), would it be infinitely good?
The idea behind FSs is to make new ones as storage increases.
In the 80s when no one did use volumes over 1TiB making a 2TiB volume limit was understandable.
These limits are made because creating to big limits in the 80s (1 YiB max volume size) will cost to much overhead from the file system.

This 2TiB limit is crippling today

I don't give a shit if you add a number at the end of the new name (EXT5) or invent a new name: Look at history and the FAT32 FAT16 limitations.

>If you are actually hitting the limits of ext4 you probably shouldn't be using ext4 anyway I don't understand the problem here.
Where is my EXT5.
>If you are actually hitting
Will hit in the next upgrade.
Can use EXT4 on life support and tricks after that point.

You pretty much did more to protect your stuff than OP. In all seriousness though, a robust and cheap way to store important files in case your main device(s) fail(s) is to buy a bunch of DVD-Rs and burn your data to them with additional redundancy (look up Parchive) to account for eventual disk rot. In proper conditions disks last a good while. It's generally a good practice to store data on several different kinds of media.

so use btrfs

And what has this factoid in common with the fact that large files over 16TiB will be impossible on EXT4.

>Where is my EXT5.
You'd use XFS.

Or even just ceph/moosefs... which ultimately work with smaller chunks / objects and suit fully distributed storage better.

>16TiB (HARD LIMIT)
Yup.

> 50TiB (Soft limit)
Fuck off. It's 1EiB.

No one cares whether RedHat will certify your setup or not if you use make a filesystem that big. Your self-built computer with even just a 2TB filesystem isn't actually certified by Microsoft either [and that $5 maximum warranty you get is more of an insult than anything else].

>being this retarded and stuck on the first sentence
Read >You're saying a spoon is shit just because it can't be used as a screwdriver. Stop making dumb complaints.
Yes, >TB files are irrelevant for desktops and servers. No regular user stores or needs that much data in a single file. An average user has 1.5TB of storage and since we're moving from HDD to SSD this won't change due to costs.
>Do you have files over 2GiB?
Retarded counterargument. FAT32 is outdated, EXT4 is modern and basically one of the standards now. The fact is that it works for everyone but 1 user. So it's his fault. Why not use NTFS when it's "so much better"?

Are you dumb? Who the fuck would watch a RAW video or store it on their main storage? People are still watching 2GB sized 1080p movies and are fine with it. There won't be a leap above 8K in at least a decade so even if you're hoarding 8K videos you'll need 500 movies to reach 16TB. I mean, 768p along with 1080p are still the two most used display sizes. If you're arguing for TVs then again, irrelevant. You'll either use external storage or stream it either locally or remotely.

>Why is there no EXT5 with larger file capacities?
Because it's irrelevant for EVERYONE but (you).

NTFS is garbage. Read again.

>This 2TiB limit is crippling today
Said nobody. And workarounds for this are trivial and if you really needed this feature you'd already figure out what to do instead of crying here.

>Where is my EXT5.
Make your own. I doubt you have enough brains to do that though.

Stop shitposting, faggot.

user is saying that no-one is paying developers to get bigger than 16TiB files on EXT4 and anybody who needs that is using some other file system.

I bet they are actually paying developers to keep EXT4 as stable as possible and to keep super high standards for any feature changes.

>Where is my EXT5
They decided to make btrfs instead since they wanted to introduce backwards incompatible features.

You do realize that this 16TB size limit only applies to 4k block sizes right? And NTFS is affected by the same limitation, and for both filesystems to get larger than the default sizes, you can simply use a larger block size, right?

IIRC, the max block size is 4MB, so theoretically the max file size is actually 16PB.

>the way it was explained to me
You where told lies.
I was once told linux dose not need defragmenting because
"linux is like a woman and because woman are better at organizing it never needs to defragment"
>ext4 never needs defragged
More like EXT4 has practically no tools for defragmenting so the unix fanboys will say it dose not need defragmentation.
Every file system fragments, there are scenarios where its impossible to not have fragmentation.
>it saves the entire thing together,
Nonsense are you talking about COW FSs?
en.wikipedia.org/wiki/Copy-on-write
EXT4 is not a COW FS.

And even COW FSs can get fragmentation depending on the use scenario.

Fragmentation is not that bad however its something that is nice to have the ability to preform.

>this kills the ssd.
All normie tasks will be practically the same on all FSs. So there is no reason to fear COW FSs killing a SSD.

>NTFS is garbage. Read again.
Is some strange philosophizing if a infinite capacity FS is infinitely good.
Can not prove if NTFS is bad or good.

Currently you'd need to patch to use larger block sizes than 4096. So basically he's right at this point.

>they are doing everything they can to get people onto Windows
Why wouldn't they?

>being this retarded and stuck on the first sentence
literally inserting meaningless sentences into a discussion.
>You're saying a spoon is shit just because it can't be used as a screwdriver. Stop making dumb complaints.
Literally using terrible metaphors.
So what is the knife and spoon FS then? Where is my Hammer FS? Every journalist FS is more like a brand of car and NTFS has far greater capacity then EXT4.
Also FSs are abandons and replaced by next versions with increased their capacity see:
EXT 1
EXT 2
EXT 3
EXT 4

Also all the switching
FAT 8
FAT 16
FAT 32

>Retarded counterargument.
its literally a extrapolation of your logic.
You literally wrote
>Limitations exist for a reason. To prevent retards from doing something they shouldn't.
Retarded shit like having files larger then what FAT 16 allows?

why THE FUCK would you store uncompressed 8K video ? If you want to edit it you know you can compress/decompress on the fly right?
Even if you did want to store it uncompressed why THE FUCK can't you split uncompressed 8K video to chunks less than 16TB ?
The only possible explanation for this thread is that you are windows retard shill, Even if that is the case you can still mount your trash NTFS in linux you sperg and it will work with your cancerous 16TB masturbation folder, and you can defragment it till the end of your days so stop wasting our time

Well the limit seems more than enough for your typical desktop user. I'm sure there's better fs suited for your specific need for huge files

>See actual documentation
>access.redhat.com/solutions/1532

Xfs has 100tb file size limit it literally says it on the site u posted

>FAT32 is outdated, EXT4 is modern
Says who?
You wrote
>Limitations exist for a reason. To prevent retards from doing something they shouldn't.
This sentence says all imitations in FSs exist for a reason and its to prevent you from doing stupid shit.

You begin to understand why your comment was retarded on all levels?

You can split 32TB files in 4GB chunks and work with FAT32 if you are a non autistic person it will always be a bad idea to have huge files no matter what the filesystem and its limitations..

>Why not use NTFS when it's "so much better"?
And here we go.
This is why clowns like you are he cancer of this GPL community.
You pretend "linux" is magically perfect and never improve it.
Whoever questions this dogma is shunned.

I understated you meant it to be some insult however I answer your retarded insult+inquiry.
>Why not use NTFS
1) I want to get away from MS
2) NTFS collides with PosixI can not use file name characters I can on Posix like ?
3) Using some compatibility layer to run NTFS is a bad idea because it can introduce problems see the filename example.

dont know man, this isnt an area of interest of mine. I was just repeating what I was told so people can tell me how wrong it is.
basically was told that ever save, ext4 relocates the entire file to the end so its all consecutive, making a trillion writes for every file that gets updated would destroy a drive and slow down performance beyond comprehension, tlc ssds wouldnt last a week.

the human brain (even the double digit iq ones) is capable of maintaining several petabytes of data. 350 to be exact. After that, all the new stuff you learn starts replasing the old sruff.