ReiserFS v4 released

Sunner

Elite Member
Oct 9, 1999
11,641
0
76
Let's see if the developers can stabilize it and make something great out of it.
Still not trusting v3 enough to put it on any kind of critical server, perhaps 4 will be their lucky number :)

Namesys
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
Should be good.

It's very fast, and you will never get to a point were you'll be left with a corrupt filing system, as long as the hardware holds up.

FSCK? Why? There is now no need for it.

Does it do journalling? Hell, no. There is no point.

Your filing system will never reach a inchoherent state. Your files are either written, or not. Your files are either moved, or not. Your files are either fully copied, or not. Never will their be a file half written. Never while you ever experiance problem during a file move will you end up with a missing or corrupt file.

Power outages? No sweat. As long as you save your working files often, then you will never have anything to phear, as long as your hardware is good.

Fully atomic filing system. All file movements/manipulation are accomplished completely in one step.

It's been attempted before, but nobody before Reiser could do it AND make it FAST.

Dancing Trees. Who'da thunk?

Now hopefully it will quickly get to "stable". Hopefully it's possible it will be stable. This could possibly be the filing system to end all filing systems. Make NTFS look like Fat16. :p


Now all we need is non-volatile RAM for our PC's and I'd be happy.
 

silverpig

Lifer
Jul 29, 2001
27,703
12
81
Well I just made a reiser4 partition and have it mounted as /home. Works good so far.
 

n0cmonkey

Elite Member
Jun 10, 2001
42,936
1
0
Originally posted by: Nothinman
There there, think positive

Yea, because their past work has been so spectacular...

If it does work out though, it could help revolutionize the way filesystems work.

Not like other filesystems have never had problems. :p

I just wish they would release the code under a different license (2/3 point BSD license), so it could be a bit easier getting it spread around.
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
The way I figure it, the best thing to do is never try to do anything new. Because if we try new stuff it could fail because it's not as well proven as something that has been being used for years. Like Fat32. :p

ReiserFS 4 is some realy cool stuff. If it weren't for the wear and tear on the hardware, a perfectly acceptable way to try to save a file from undeleting would be to immediately pull the plug on the computer and hope that you cut power before the file operation finishes.

Imagine a file system that works the way people naturally expect a file system to work.

No journalling. Asyncronous file operations are fast and safe. It's extensable and flexible.

Some weird stuff, though. For example making a file a file AND a directory. Kind or what like Apple's HFS+ does, or some of NTFS's semi-undocumented features]http://patriot.net/~carvdawg/docs/dark_side.html[/L]. (Nice tricks for virus/worm writers though.)

Say you have a file named "helloworld.txt"

you can go:
ls -l helloworld.txt

But you could also go:
ls -l hellowworld.txt/
and see the files inside helloworld.txt that contains metadata, such as ACL's or icons or whatever you want. Much like what is used in OS X. However the way they are doing it would be backward compatable with older apps. You wouldn't have to modify ls to utilize it.

I am thinking that this would be usefull for creating file sharing software to be more compatable with OS X's way of doing metadata, or Windows XP SP2 NTFS's way of storing extra security data in the alternative data streams (that's what is used when download a file and it say's "it's not from a trusted source" type things.). Stuff like file sharing with ReiserFS4 aware SAMBA could benifit from it and be able to more well integrate with other OSes. Or something like that. (all quite a bit above my head).

Although I understand the lack of incentive to add such features to programs, but it could lead to other developements, I suppose.
 

djdrastic

Senior member
Dec 4, 2002
441
0
0
Do I just need to upgrade my kernel for support ? I havent seen any weird modules for it yet ?
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
Yes, a kernel upgrade would probably be in order. Maybe even a kernel patch. I don't think that it has been accepted into the mainstream kernel yet and even if it was, I considure it mostly experimental at this point.

At this point, you have to know what your doing in order to play around with it.

from here
WARNING: Reiser4 has just been released. Namesys has the most stringent QA process of any Linux maintainer, a firm belief that no user should hit a bug that developer testing could find first if the effort is made, a by now grand archive of filesystem tests, and no one on our mailing list can make it crash any more, and all of this is why it took us so long to release it, but until a program has seen a few million real users you should not use it for a mission critical server. For that, use ReiserFS V3 for now, as V3 has millions of users and goes many months between bug reports (and the bug reports we do get relate to tweaks someone outside Namesys added, not core code.).

installation overview

You would have to use a experimental version of 2.6.8.1 in order to get Reiser4 working.

(reiser4 and resierfs v3 are 2 completely different filing systems, BTW. Each reiser filing system is unique in itself. Unlike, for instance, ext2 and ext3. You can take a ext2 filing system and upgrade it to a ext3.)
 

Sunner

Elite Member
Oct 9, 1999
11,641
0
76
Originally posted by: djdrastic
Do I just need to upgrade my kernel for support ? I havent seen any weird modules for it yet ?

Reiser4 will be included in the main kernel tree sooner or later, but for now, you'll need to download it in patch form and apply it to your tree.
I think Andrew Morton included in his tree as well(the -mm tree).

Andrew's tree is available at http://www.kernel.org, and the patches for the vanilla 2.6 tree are available here.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
If it does work out though, it could help revolutionize the way filesystems work.

I definately think it could be great, Hans has some really good ideas. The problem is in the implementation and development process though, suffice it to say that I don't trust them.
 

Sunner

Elite Member
Oct 9, 1999
11,641
0
76
Originally posted by: Nothinman
If it does work out though, it could help revolutionize the way filesystems work.

I definately think it could be great, Hans has some really good ideas. The problem is in the implementation and development process though, suffice it to say that I don't trust them.

One could hope they've learned from Reiser3 and will improve this process when maintaining Resier4.

Time will tell, I'll steer clear for a while before I try it out, but hopefully, it can evolve into a kickass little filesystem.
 

n0cmonkey

Elite Member
Jun 10, 2001
42,936
1
0
Originally posted by: Nothinman
If it does work out though, it could help revolutionize the way filesystems work.

I definately think it could be great, Hans has some really good ideas. The problem is in the implementation and development process though, suffice it to say that I don't trust them.

I can't say I care much for the development process either, and I wouldn't trust it as much as I'll trust other filesystems. But there are always backups. :D
 

sciencewhiz

Diamond Member
Jun 30, 2000
5,885
8
81
Hans is like Linus, he gets bored with maintenance. If someone steps up and is serious about maintaining Reiser4, it will go a long way. If not, it will be just like Reiser3, just a hobbyists FS.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
Hans is worse though, Linus atleast makes effort to maintain compatibility and not break things unnecessarily.