Your chosen file system and why you use it...
 

After seeing people mention a couple of file systems in the pulse audio thread, I thought it might be interesting to see what everyone's using.

I stick with ext4 for my / and /home mainly for the decent trade off between sequential and random reads/writes. For my media and games drive I use XFS because I find it fantastic with the reads.. then again that may be entirely in my head so feel free to correct me :D

What do you blokes use and why?

Cheers.

I use ext4 for everything for the K.I.S.S factor. Any slight performance gains I might get from mixing and matching filesystem types simply aren't worth the added management complexity. YMMV.

ext4, it's the default, it works, and I haven't bothered trying any alternative.

While I've read good things about BTRFS especially in server/workstation environments, ext4 is just easy and decent so for now and awhile, that's the one.

EXT4.... finally.
I had resisted that, sticking with EXT3 until about 2013, for REASONS!
I resist change, but I am finally sold on EXT4

Maybe have a look to some of the previous answers...
https://www.linuxquestions.org/quest...em-4175526093/
https://www.linuxquestions.org/quest...em-4175598044/
https://www.linuxquestions.org/quest...sd-4175550779/
https://www.linuxquestions.org/quest...on-4175428850/
https://www.linuxquestions.org/quest...hy-4175479794/

BTRFS

Being able to make cheap snapshots and send backups of the live system is amazing.

Being able to roll back (or just reboot into the snapshot!) a set of updates that broke some third party app you have installed when you got not time to troubleshoot is priceless. You can even do entire slackware upgrades!

Doing easy online incremental backups is also wonderful. I can keep months worth of history and restore to any given point (I have tested restoring to VMs) without a lot of storage. Also indispensable if ever you land in a situation when you don't know when something broke or was compromised.

Oh well the 'search for similar threads' function is not working very well then, because I used it before posting and none of them were returned. Anyway, the last Slackware thread was from 2012 - I think 5 years is reasonable gap, a lot changes in that time and surely you've heard the old addage of not to revive old threads but start a new one ;)

Cheers for the other replies :D

Edit to add: oops - one of the slackware threads was from 2013.. still a long time.

Just plain ol' ext4.

I have long (since 2005) preferred xfs for its performant handling of very large subdirectories and excellence in reliably self-recovering from unclean unmounts, but have started using ext4 in recent years as well.

Nowadays my systems are all using SSDs for their boot/root disks and spinning disks for bulk storage. I have opted to put ext4 on the SSDs and xfs on the spinners. The ext4 dev team has kept more on top of TRIM support than the xfs team, and by my benchmarks ext4 handles lockfile contention about 30% more rapidly than xfs (which is an issue for me because of specific software I use which depends on the https://metacpan.org/pod/File::Valet lockfile implementation to mutex-protect shared logfile appends).

The bulk storage disks retain xfs because of its aforementioned high scalability and reliability. I can have millions of files in a single subdirectory without suffering severe readdir() or open() performance degradation. Maybe ext4 has gotten better at this lately, but last I checked (2011'ish) it did not.

When I need "next generation" filesystem features, I use md and GlusterFS instead of zfs or btrfs. It has the advantage of RAIDing storage across different machines in a network, not just within a single computer. It works great under Slackware with no need for special kernel support. Configure, make, make install and it's good to go on a generic Slackware system.

GlusterFS implementation is all userspace, but performs fairly well nonetheless -- it can saturate a gigabit ethernet link with reads or writes unless running on a very underpowered processor. It provides a cifs export interface, so Windows and MacOS users can access its filesystems over the network, and gives me RAID6-like redundancy at the host level, so I can yank a data server off the rack for repairs without interrupting service (or two, if I feel like living dangerously).

Since GlusterFS "bricks" are based on subdirectories in an existing filesystem, several filesystem layers are involved. For my configuration I use md locally to RAID5 (or RAID50) the host's disks, format the array xfs, create my brick directories in mountpoint /var/bulk/, and configure GlusterFS to turn the bricks into network-mountable RAID6-like volumes.

ext4, played with BTRFS on openSUSE and Mageia, had problems with it on both...ext4 (and before that 2 & 3) has never failed me....oh God...must knock on wood now...knock knock... :)

Ext4 because it's the Linux file system........ ;)

Ext4. I stick with the default on my laptop and work stations. It works well for me.

ext4.

Generally, when I install a distro, I accept whatever the default is for that distro. Part of it is laziness, the other part is that I reckon the maintainers have Reasons.

Same here. I know ext4 is super stable, super well supported, and while you might gain some performance in certain circumstances with other file systems, I don't really need that added performance desperately, so I just stick with EXT4.