NTFS Driver Lands Some Late Feature Enhancements For Linux 6.12

[erniv2]

Have to push the forum count why does a late ntfs3 pull only get a slight slap and bcachefs gets the hole shitstorm ???? Btw. dont feed the troll

[NM64]

Because most people dont live in perfectly linux-siloed worlds. While not an everyday thing i interact with my own windows boot, and those of my friends and family often enough that writing is needed. The most recent FAT variants that dont have that 4gb file limit are fine, but NTFS is also better with features that matter

I use a slew of older smaller-ish SATA SSDs as essentially large floppy disks (can be connected via SATA, eSATA, USB, whatever a given situation needs), and one thing I found is that exFAT on Linux regardless of disk or distro or kernel version was much less stable and liked to disconnect if you were trying to read something like 15 large files at the same time while NTFS on Linux (well, the FUSE driver rather than this Paragon driver) handled it without a hitch.

I was particularly running into this is mainly when reading a bunch of checksum files to verify data integrity of some larger archives and/or backups. Because it's an SSD and we've now many-thread CPUs, doing so many concurrently was a complete non-issue as long as the file system wouldn't bork out.

(and to clarify, it's not an inherent exFAT issue as even Windows 7 would handle the very same task without issue)

I also found out the hard way that exFAT doesn't support symbolic links, and I really like to use them quite a lot for all kinds of things (most commonly when I manually directly modify a file for a program, I'll rename the existing file to BAK and then use a symlink to point to a folder in the parent directory that is a replica of the program's folder arrangement but only containing my custom and/or modified files)


And on a similar note, the MediCat live rescue USB requires using NTFS rather than something like exFAT, and I'm forced to use such a thing in order to create checksum files recursively via HashCheck Shell Extension even on EXT4 disks because, last I checked a couple of years ago, there's no GUI-based native Linux software that can do so (for small amounts of files I get away with using ExactFile running under WINE, but it borks out and/or takes multiple hours for something that HashCheck shell Extension can do in like 30 minutes without issue).

(protip about MediCat: despite being a 20+GB download, the actual required files for the "Mini Windows 10" to actually function amount to only around 1-2GB, the bare minimum only being the Mini Windows 10.wim file along with the entire root://ventoy and root://system folder.)

[pWe00Iri3e7Z9lHOX2Qx]

Isn't the ntfs3 driver still considered experimental and not enabled by default?

I'll be the odd man out and say I wish it was the default in more distros. I've had several bouts of weird behavior with the ntfs-3g driver where ntfs3 worked properly. But just using ntfs3 at all in some distros is a pain because it might be blacklisted, mkfs and other utils symlink to the ntfs-3g driver, etc.

[User29]

this whole ntfs thingy always sounded very dangerous.

[Alexmitter]

I wonder how to make Konstantin address this: https://bugzilla.kernel.org/show_bug.cgi?id=219123

I've written to the mailing list three times, I've emailed him personally twice, I've created a bug report. Nope. No one cares.

Birdie, maybe just maybe no one gives a F about what you say. Its not like you aren't well known beyond this board.

[PuckPoltergeist]

I wonder how to make Konstantin address this: https://bugzilla.kernel.org/show_bug.cgi?id=219123

I've written to the mailing list three times, I've emailed him personally twice, I've created a bug report. Nope. No one cares.

I'm pretty sure there was an answer that this behavior is intended and won't be changed. I've encountered this errors myself and indeed after checking the filesystem from windows, the errors are gone. So the real bug is, that ntfsfix doesn't notice and can't fix it. About the bugreport, this should be closed as won't fix/works.

[varikonniemi]

Have to push the forum count why does a late ntfs3 pull only get a slight slap and bcachefs gets the hole shitstorm ???? Btw. dont feed the troll

because bcachefs is a repeat offender

[gnarlin]

You can use ntfsfix.

ntfsfix v2022.10.3 (libntfs-3g)

Usage: ntfsfix [options] device
Attempt to fix an NTFS partition.

-b, --clear-bad-sectors Clear the bad sector list
-d, --clear-dirty Clear the volume dirty flag
-h, --help Display this help
-n, --no-action Do not write anything
-V, --version Display version information

For example: ntfsfix /dev/hda6

Though not "native" ntfs it works good.

No, I can not. You need to read up a little more about the situation.
Ntfsfix does very little. It can only clear the dirty flag and bad sector list.
It does NOT actually fix problems on corrupted ntfs partitions like chkdsk can.
Paragon does have an actual proprietary linux utility for fixing ntfs paritions/volumes and that's what I'm currently using, but I care about Free software and if paragon released the source code for that utility we would finally have a propper fsck.ntfs in our package managers, ready to be installed and used when needed. Inspected, improved and bug fixed over time like all Free software can. Paragon software have already done a lot for us by releasing and committing to maintaining the Free version of their ntfs kernel driver and I'm incredibly thankful, but that fsck.ntfs puzzle piece remains missing and has been for all my life. I rarely use Windows at home (don't have a permanent install anywhere) but I use it when I must for some project or other. At work I have a dual boot workstation and need to regularly use various Windows utilities and applications, not because I want to but because I have to. Every small piece that enables me to use proprietary software less is something I'll celebrate (with a small muffin and a glass of milk). Among my duties at work is data recovery and pc repair (including fixing people's borked Windows installs) and some utilities required to fullfill those duties require proprietary Windows software, whether I like it or not.
On a related note, Gparted, while competent, has a terrible layout compared with the proprietary Minitool, which gives you a good overview over all the drives at ones. Good ideas are worth stealing and if Gparted had the same disk overview that Minitool had it would be a step forward. Another small missing piece is that Minitool can clone drives to smaller drives but Clonezilla and Gparted cannot. It's one of the reasons I am forced to use Minitool in Windows for some disk work. Anyway, I didn't mean to ramble but here I am. Please don't tell people that ntfsfix is equivalent to chkdsk.

[mdedetrich]

Because most people dont live in perfectly linux-siloed worlds. While not an everyday thing i interact with my own windows boot, and those of my friends and family often enough that writing is needed. The most recent FAT variants that dont have that 4gb file limit are fine, but NTFS is also better with features that matter

You should be using exFat for thumb drives/external USB drives, it can be mounted and written to by every single modern OS (mac, linux, windows, android etc etc) and it doesn't have FAT's 4gb file size limit.

[mdedetrich]

Have to push the forum count why does a late ntfs3 pull only get a slight slap and bcachefs gets the hole shitstorm ???? Btw. dont feed the troll

And its even worse/later than bcachefs transgressions as well.