My first steps in CachyOS…
Because I’m running dualboot with windows11 there is a partition I share on both systems. G-Gemini (GG) told me putting NTFS on this drive would be best to go. And I have an additional btrfs-drive for games on linux.
Then GG told me to add this to my fstab before mounting:

# BTRFS-Partitionen
UUID=xxx   /mnt/Lin_Games   btrfs   defaults,nofail,compress=zstd:3                     0 0
# NTFS-Partitionen mit schnellem ntfs3-Treiber
UUID=yyy   /mnt/Lager_Lin    ntfs3   nofail,uid=1000,gid=1000,iocharset=utf8,umask=000   0 0 

()
Yes, it works so far … but finaly I’d like a human-opinion on this.

What do you guys think about my fstab-additions? Are there any further recommentadions to the mounting??

These LLMS are often not good about particulars .. so not so good with nitty-gritty of tech .. or linux especially .. and especially something like Arch/Cachy which is niche even among linux.

So for example they probably do not realize we have access to the newer ntfsplus ntfs module and that its very preferable for multiple reasons (data reliability .. also the first and only example of actually being able to repair NTFS without windoze). As long as you are using kernel 7+ (currently default).

Also why not using defaults?
Why using UID OR GID?
(NTFS does not use permissions because a a proprietary filesystem by and for msoft)
Why umask?
Also probably should be 0 2 at the end so that repairs can occur.

Though some of those could be a conscious choice for some reason.

But at first glance it looks like the only parts I agree with are nofail and the use of UUID.

For an example of what I would generally recommend;

Thx for your advice!
Sure, the whole LLM-thing is not about beeing good … it’s only about beeing ‘good enough’ And for me the LLM did a job good enough to make my Cachy running games and letting me know that drives need to be mounted plus “hints” for the mounting… all that in my native language
Without the LLM I wouldn’t even know what to ask here often times.
Of course these “hints“ need to be discussed, verified and surely can get buffed … all by real human experts. -As in this case! So, regarding the questions why Gemini did set various options for me:

• ntfs3 –> yeah, Gemini confessed not beeing fully up2date on this.

• UID + GID –> I would not be able to copy new movies onto the drive, rename files, or delete old videos without entering my admin password in Dolphin every single time. uid+gid=1000 should ensure that the partition belongs to me, allowing me to manage files without annoying password prompts.

• umask –> oh, that was a relict from trying to use this ntfs-drive for steam-games … bad Idea - I know NOW^^ (the LLM did not much advice against that in first place)

• 0 0 at the end –> Should prevent troubles for noobs like me with dualbooting windows (10+11)

And what do you think about Geminis latest suggestions for me (when told that I wanna use it as storage for movies and stuff):

• dmask=000–> full permissions for all folders (creating new folders, opening them, deleting them).

• fmask=111–> A movie stays a movie, a picture stays a picture. I could create, watch, and delete them, but they can no longer be accidentally launched as a program.

If you could comment all or some of these points, I would appreciate very much.
*edit* And how should I do for the btrfs-partition??

it’s about offering you s slot machine that might be good enough or might trash your machine while discomnecting you from real documentation and humans.

And now that my machine is working (also with AI’s help) and we are connected here …
… You do agree with all the stated points from my last post?

Sure thing. Here goes;

Not usually true, especially if you have a working fstab line.
I would think you are either speaking of the mount itself or some other weirdness.
See ex;

How is this preventing troubles?
You are just making sure it never runs FSCK automatically.
If you were using any of the prior reverse-engineered modules then they simply would not have any fsck that works with ntfs anyways. The 0 in those cases is ‘right’ I guess but also happens to be redundant.
With the newer module it does support a functional fsck so you could run fsck and it actually mean something.
So I would think you want that 2 because its not root but we do not want to disable filesystem health checking.

Again not normally applicable to an ntfs part.
If it is we still probably dont need or want some special ‘everything’ perms.

Thats not what that does.
That would mean everyone including group and owner have permissions of 1 .. in octal that means they have no rights to read or write .. but can execute? This is an odd option and does not do what it claims.

You mean the fstab file?
I do not use btrfs but options something like the following;

defaults,nofail,x-systemd.device-timeout=5,noatime 0 2

Though some people say they want lower compression and more speed so they add that;

defaults,nofail,x-systemd.device-timeout=5,noatime,compress=zstd:1 0 2

(compression default is zstd:3 )

well…no not really. there’s a lot of straight-up bullshit.

For me as a total linux-noob there’s one trifle left:
Why is it mounted under /media/secondary (and not /mnt/)? Is it just personal preference or is there any advantage??

And also two biger things:

ntfs defaults,nofail,x-systemd.device-timeout=5,noatime 0 2
btrfs defaults,nofail,x-systemd.device-timeout=5,noatime,compress=zstd:1 0 2
… Where I could have found this information (and choice) on my own?

And: Why dont you use btrfs??

Only an example.
Mount it wherever you like .. though some paths do have certain properties.

Maybe here;

And more pointedly heres;

As well as man (manual) pages for any of the specific filesystem drivers/modules, ex;

It has had serious reliability problems.
Those are mostly sorted now… though if your data was quite serious and you were using something like RAID then you might want to make sure you have copies.
With EXT4 I never needed to keep backups… and besides being iron-solid (more than ‘rock’ but maybe not a diamond?) its also pretty darn quick.

So long as I do not need any of the fancy features then EXT4 remains a great choice .. maybe XFS if dealing with large files all the time.

That’s truly a bummer for something you’d choose partly for data integrity functionality.

And also too bad the zfs license means it’ll always be outside the kernel and thus not receive the same “easy to use accessory program” (even if support on cachy is excellent) development attention that btrfs does.

When filesystems cause loss of data it tends to leave a scar.

Though I hope it does not come off as too scaremonger-y because btrfs is used all over the place and has matured a fair bit.

( But I am still not going to call it “butter-eff-ess” )

ZFS is interesting especially if it fits your use-case. Another thing about it is its always been memory-hungry (maybe do not use it anywhere that has less than 8gb of RAM or where memory is otherwise at a premium).

I use ZFS largely due to a bitrot incident 20 years ago.

I trust btrfs for other people’s data

It’s not quite as memory-hungry as reputed, and has had a few consumery flexible features grafted on recently but it still has heritage as a filesystem designed to be preplanned without hardware budget as a driving concern.