Benchmarking The Experimental Bcachefs File-System Against Btrfs, EXT4, F2FS, XFS & ZFS

[k1e0x]

XFS held up well in these. SGI clearly had some really smart people working for them. From these benchmarks, it seems like XFS might be a good choice for server work-loads with lots of random activity.

What I don't understand is why ZFS still has such a dedicated following. I'm not claiming they are wrong, I just am confused. If anyone should be backing ZFS, I think it should be Oracle and yet Oracle Linux seems to prefer btrfs. Oracle Linux does not officially offer ZFS support at all. In terms of LInux file systems, Oracle seems to be putting their resources into both btrfs and OCFS2 while also being willing to supply both under the GPL making the license compatible with the Linux kernel. If we are going to continue to look at ZFS as an option, can someone please get Oracle on board with it?

I think it's troll but.. to answer your question it's very simple. Because it's good.

It's the most reliable filesystem you can use and it makes managing that data easy with it's features and flexibility. (It's also not slow, but it is slow in random reads/writes)

Oracle has nothing to do with Open Source ZFS. ZoL or BSD's ZFS.
It's like saying "I don't understand why people use LibreOffice, Oracle doesn't care about Open/Star Office!"

Who would willingly use Oracle software anyhow? You? lol I don't *want* them on board. Their software sucks and they ruin almost every open source project they touch. (cough mysql) Companies that back Open Source ZFS are Delphix, Datto, Nexenta, Canonical, Joyent, DDN and IxSystems with *thankfully* no support from Oracle at all.

The more you know :: whistles tune ::

[jpg44]

A lot of focus on SSDs here. SSDs are fine. But isn't it the case many users still have a HDD because they are simply less expensive. Need to also think about the people who dont have deep pockets and the most top of the line hardware.

[nomadewolf]

already awful results for fs which doesn't even have featureset of competition. adding features will only make it worse. but feel free to organize church

The developer clearly stated that, so far, the goal was to get the FS stable. Performance wasn't considered at all.
Also, the extra features will only impact performance IF they are used. That's true for every (decent) FS. so i don't get your logic...
So, from now on, we can only expect performance improvements (a part from the regressions every piece of software always gets).

[pal666]

i would like to see a "fragmentation count" benchmark.
still disappointed by btrfs with huge vm-disks - and the performance degrade after a while because of that

disable cow on vm disks ffs

[pal666]

Already solid results for an FS that's not even mainlained.

already awful results for fs which doesn't even have featureset of competition. adding features will only make it worse. but feel free to organize church

[pal666]

lol, featureless fs took last place in geometric mean. no second coming this time

[ypnos]

That's because the testing was done on an Optane drive which is similar to an SSD except it has much better access times and random access times. F2FS is design specifically for drives that use solid-state storage technology.

I figured as much, but damn, look at how well it succeeds in that goal. The best/most consistent results I have seen from this filesystem so far. A match made in heaven.

[profoundWHALE]

Can we talk about the overlooked champion in this benchmark? To me it appears as if F2FS really dominates the field, being among the top in all of the benachmark categories (as shown in the radviz chart).

That's because the testing was done on an Optane drive which is similar to an SSD except it has much better access times and random access times. F2FS is design specifically for drives that use solid-state storage technology.

[ypnos]

Can we talk about the overlooked champion in this benchmark? To me it appears as if F2FS really dominates the field, being among the top in all of the benachmark categories (as shown in the radviz chart).

Apart from that bcachefs kicks ass, closely watching!

[geearf]

Thank you for the bcachefs tests!
I think it'd be nicer to add some more non-pro tests in the future.
A simple one would be mount time, btrfs can get pretty slow with partitions over Tbs and with a lot of files.
Without regular metadata defrag, some of my btrfs partitions can take more than 30s to mount... whereas they mounted under a second with ext4.