ext4 or reiser4?

kEiNsTeiN · Guru Joined: 29 Apr 2005 Posts: 361 Location: Germany

mkay, here's my ebuild for the latest e2fsprogs. obviously as patch

... but pretty straigt-forward.

there IS a livecd with all that: "Parted Magic" !

hope this helps! (using windows vista myself since a week... new laptop and all... *duck*)

Matteo Azzali · Retired Dev Joined: 23 Sep 2004 Posts: 1133

kernelOfTruth · Posted: Mon Sep 03, 2007 4:54 pm Post subject:

Btrfs ftw

it's extremely promising and I hope the developer includes compression and encryption support as well - then this will be THE filesystem
even at this early stage it's already faster than the reiser filesystems (which are the fastest from the experience & tests I've run)

i92guboj · Posted: Mon Sep 03, 2007 9:40 pm Post subject: Re: ext4 or reiser4?

Matteo Azzali · Retired Dev Joined: 23 Sep 2004 Posts: 1133

marsclic · n00b Joined: 02 Mar 2003 Posts: 74

Since I've been using Reiser4 for more than 2 years now, I can say from experience that it's very reliable. I am using it on a 4 disk raid-0 on a quad opteron. I've been experimenting with this box a lot and I've had lock-ups due to experimental kernels, over-volting and over-clocking, not to mention the storm last week that blew-away both my UPS's. Through all of this Reiser4 hasn't lost a SINGLE byte of data. I haver never experienced ANY data-loss with Reiser4. Unfortunately I cannot say this of ext3 or for that matter, Reiser3. I also have benchmarked Reiser4 against ext3 and xfs and found it to be the fastest filesystem for my particular hardware, by a long shot. Reiser4 has been released a long time ago, while ext4 is still in heavy development. I have used both the -mm and morph kernel series. So my recommendation would be to go ahead and use Reiser4 if your hardware is good enough, and make sure you have a bootable recovery CD/DVD with Reiser4 included.

Habbit · Posted: Wed Sep 05, 2007 12:13 am Post subject:

I hate to enter a possible flame war, but I must say that the probabilities of Reiser4 ever entering the mainline kernel (Linus' tree) as ReiserFS 3 did are near to zero.Why? because, amongs other reasons:

Its inventor and main developer is not available. It doesn't matter that he is in prison, or on a vacation in Hawaii, just that he is no able to support his creation.
It is not the first time that Hans Reiser has suffered from the "kid programmer syndrome": he starts something, leaves it in a semi-stable state (reiserfs) and then moves on to his next "grand creation". This is not necessarily a bad thing (I can't track count of how many "grand projects" I have unfinished), as Linus did it himself - but he returned to support and manage his creation. Support for ResierFS has been nearly discontinued, and even bugfixes are no longer being introduced without extreme resistance by the developers
Reiser4 is based on recent technology. This is both good (there _could_ be a lot of room for space and improvement, _if_ the concept is good) and bad (it is untried relative to other FSs whose base technologies have been around for much longer)

This does not imply that Reiser4 is not a good filesystem, only that the maintainers of the mainline kernel consider its developers undependable on. Which means that you will always be relying on patched kernel sources - but this is not new, most of us use gentoo-sources, so distros could include Reiser4 with their patches.

This being said, I have 3 boxes: my secondary box and my laptop use reiserfs as a high-performing root partition, but my main box has ext3: why? stability. It's not the first time I've seen reiserfs go down in flames (and rebuilding the tree usually discard some 33% of files :roll:

). I hope Reiser4 improves in this aspect, which, as I said, will no longed be improved in ReiserFS 3 because the developers have completely moved to reiser4 (as a comparison, the Linux 2.4 kernel is still being mantained for bugfixes)
_________________

Matteo Azzali · Retired Dev Joined: 23 Sep 2004 Posts: 1133

@Habbit: there's one more reason. After claiming that Reiser4 was ready for insertion into kernel
for 3 years, and refusing any critic on how the code was written or the non-compliance of the plugin system,
Hans Reiser declared that there were bugs and things to fix and that in some months there would have been
Reiser 4.1 to fix everything (in which case someone to review his kernel-compliant code whould have "spawned").

In july 2006 he talked about "a month or two for reiser 4.1"(search google if you're skeptics). In october instead,
it was still not ready and instead happened that "accident" (the arrest, I don't have anything against Reiser and I hope he'll be
free if innocent), Hans Reiser sold the Namesys, and sadly actually still nothing about Reiser 4.1 .

Other than this there's also a little issue about bugfixing policy (read it carefully, please).

Chanches that Reiser 4 / 4.1 will ever enter the kernel are actually under 1%, if someone doesn't start to clean
and rewrite that code (to be written within kernel quality standards, support the VFS plugins/modules system,etc.etc.)
and then grants some kind of support/bugfixing (in the GPL meaning of support/bugfixing, no tricks on definitions).
_________________
Every day a new distro comes to birth. Every day a distro "eats" another.
If you're born distro, no matter what, start to run.
---- http://www.linuxprinting.org/ ---- http://tuxmobil.org/