ext4

So before those updates happen, I have a backup of all my files right there that can be easily shifted back to.

Now, if I wanted to make this go faster and I was smarter, I would have done this in BetterFS instead of EXT4 because EXT4 requires rsync where BetterFS is atomic, so it would happen instantly.

If you don't know what I just said, just know BTFS or BTRFS is better than EXT4, especially when it comes to backups when configured properly.

But the configuration of BetterFS is a little bit more complicated because it uses subvolumes.

But I digress.

And Windows has been using it for 20 plus years, where Linux has seen so many more improvements over the years, where you have better FS.

You have, EXT4 is pretty old these days, but it's still newer than NTFS.

And yeah, XFS, there's a whole bunch of different file systems to choose in Linux.

Typically, all of them will be better than NTFS.

Project quotas.

Ext4 supports them.

You turn the feature on when you make the file system or with tune2fs, mark a directory with a project ID and set limits per project.

A directory can't grow past its allowance, and there's no inner file system, no loop, no double journal.

And for encryption without a container.

Fscript.

Per directory encryption built into Ext4, F2FS, and a few others.

Unprivileged users can use it, and there's no mount step.

If, if the device name starts with /dev/mapper, mapper, it's on a LVM or Lux, uh, setup.

So if secure boot is on and your /boot isn't, you know, basic EX- EXT4 or Fat, you basically have three choices.

You c- you can move boot to a, you know, simple, uh, partition that has its file system with one of the compatible ones, you know, EXT4 or Fat, you can turn off secure boot, or you can stay on 26.04 LTS.

So- Or [laughs] or you can leave Ubuntu.

[laughs] Or, or that is a possibility.

You wrote, like, 50 other valentries after that, and then something in the middle got corrupted, right? So in, in some cases, like, if you think about it in a fundamental way, what is happening here is that the recovery code cannot reason whether this particular checksum mismatch was caused because of a crash or because of a disk corruption, because they appear as the same symptoms to the recovery code, right? Now, this was a very interesting finding for us when we were doing this work.

So we went back to ext4's journaling code, uh, which also has a very similar sort of construction.

You have the journal, and then the journal entries are, uh, checksum.

So we will, we will be, like, poring through the source code of ext4, and we found this place where a developer had, had commented something like, "Oh, there is a checksum mismatch in this entry.

We cannot tell whether we crashed when we were writing this or it was just a random corruption.

We are just going to truncate the log," right? Or something or something to that effect, right? And, um, uh, that's what exactly the systems also did.