POLL: Which file system in 2022. ext4, btrfs, xfs, zfs, f2fs (choose 1)

Yes, I knew.

No, I am using Btrfs, Ext4 and ZFS in my device with 4 different disks.
My plan was already ZFS mirror for my archives, my second backup via restic and sharing like NAS.

I read for the first time that zfs-mirror cannot be imported.

1 Like

Different filesystems are designed for different purposes, there’s always trade-offs. BTRFS’s “reliability” is in the metadata and snapshot capability, which wouldn’t be helpful in a server environment but is fine for a workstation or laptop. On a server I’d want something easily extensible (like XFS) for a hardware RAID, or ZFS if I’m doing some sort of “software RAID” (more like some sort of storage-as-a-service setup). For a virtual server machine I’d DEFINITELY want XFS.

Just pick the right tool for the right job, right? I don’t get why some people feel such personal prestige when saying their choice of filesystem is the right one for all. Life doesn’t work that way.

Now OSes, those do work that way, because obviously EndeavourOS is the right one for everyone. :rofl:

1 Like

Same here, I’ve stopped reinstalling OS since I’ve got EndeavourOS. My experience on Ubuntu based distros was very short and not as comfortable as with EOS, where I had some little issues but I just need to downgrade some packages or wait for new updates to solve them.

1 Like

True, but it happens very rarely on ext4.

In 100% of cases where I lost ext4 filesystem, it was due to hardware failure of the drive (and, obviously, nothing is immune to hardware failure).

As long as you don’t have Windoze dual boot :stuck_out_tongue_winking_eye:

Nope
work makes me use it on the laptop they issued me though. They won’t spring for a Mac (which I don’t really like anyway).

Let’s say the BTRFS becomes totally corrupt beyond recovery. By restoring the entire system from a snapshot, would that replace any corruption with a good snapshop? Is a snapshop a complete system?

No, snapshots are part of the filesystem.

1 Like

So how to get a complete snapshop, just like dare I say on the Mac I have Carbon Copy Cloner, which I have used for decades and as long as its a good image, you can restore it over any corrpupted file system. It has happened to me once with a failed hard drive, and another time with file corruption. Carbon Copy Cloner just works. Currently I backup home to Pika/Borg. and I am so paranoid I also do a DĂ©jĂ  Dup. What to recommend?

If you want recoverably in the case of hardware failure or filesystem corruption you have a few options:

  • Take actual backups with a backup tool
  • Make an image of your disk
  • Replicate your snapshots to a different filesystem on a separate disk or a separate machine
3 Likes

Can you suggest some backup tools that do just that?
With imaging programs, would they be bootable?
Do you know of imaging packages, either GUI or CLI I can try?

There are many but my recommendation is always something backed by borg. It is easy enough to setup without a GUI using something like borgmatic. If you prefer a GUI, you can use a borg frontend like vorta.

On the other hand, replicating snapshots is actually a handy tool.

I would take either of those options over a disk image though others would disagree.

Of course, we are probably getting a little far off-topic.

1 Like

Absolutely, I will create a new topic.

1 Like

Which backup system in 2022: Timeshift, Snapper, Borg, CloneZilla, dd, btrfs send/receive, ZFS (zpool?) etc​:joy: Choose 1!

3 Likes

Aynthing on Reddit is what I would call anecdotal evidence.

If you are concerned about stability / reliability of an application resp. filesystem you should look at the bug trackers.

For zfs this is

with 477 open defects.

For btrfs this is
https://bugzilla.kernel.org/buglist.cgi?component=btrfs&limit=0&order=changeddate%20DESC%2Cpriority%2Cbug_severity&product=File%20System&query_format=advanced&resolution=---
with 674 open bugs

For ext4 this is
https://bugzilla.kernel.org/buglist.cgi?component=ext4&product=File%20System&resolution=---
with 134 open bugs.

3 Likes

Yes, there are many small bugs and few serious bugs,
that’s why I use 3 different filesystems, 4 disks and 2 different forms of backup (daily fast backup via btrbk and weekly backup via restic ) better than single filesystem and single backup. so most likely zero data loss and zero data corruption when hardware failure or corrupted upgrade system or accidentally change/delete or malware attack except burning whole PC. :wink:

Good find. Some systems are more perfect than others.
Then would you say installing brtfs on / and /home, and storage on ext4. Does that sound like a safer option?

I have not used ext4 since many years. My preferred kernel in-tree filesystem is xfs. When I started my career in IT 20+ years ago, xfs was on all the SGI workstations. I learned to love it and trust it. Several years ago then I decided to kick out all ext4 in favor of xfs.

I do not trust btrfs. I have seen too many reports about data loss and my own tests with it did not convince me. I use zfs for many years now and I frequently look into btrfs and give it a try. But zfs is the best filesystem I have ever seen so far. Especially because I use raid10 and raidz2.

Unfortunately zfs is not an in-tree kernel filesystem so I do not want to use it for ‘/’. But I use it for anything else. I boot my PCs with xfs on ‘/’ and all the rest, e.g. /home, /opt, /data, is on zfs.

To your question: If you are fine with btrfs, and your comment implies that, then there is not need for you to use ext4.

Safer in what sense? My thought process is very simple. I always want space efficient snapshots and transparent compression.

I use zfs where it makes sense and btrfs where it doesn’t.

As far as btrfs(or any other file system) reliability goes, that is what backups are for. Unlike @Kresimir, I have lost an ext4 volume. I have lost btrfs volumes on more than one occasion. I have lost hardware raid arrays. Physical disk failures occur at an alarming rate. There is no substitute for a backup.

Either use a backup solution or replicate your snapshots.

2 Likes

Actually, you should do both. Snapshots don’t help with filesystem corruption, you still need to have backups of important files within that filesystem. We do both at work, vm based snapshots AND backups of important files, both to tape (it’s not a backup unless it’s on a different host anyway). I’ve restored system snapshots that were still hosed, and ended up having to reprovision a new system and restore the files to it. It just happens.