Skip to main content
Welcome guest. | Register | Login | Post

Death of a filesystem(?)

Over the last months there were repeating news about the murder on Nina Reiser by her husband Hans Reiser, known in the community for his work on his filesystems ReiserFS and Reiser4.

In this post I don't want to deal with the murder, the process or the conviction, there are lots of news and posts about this all over the net, and in at least some cases written by people who are a lot more competent in that field than I am.

What I want to deal with here are the consequences of this "story" that may be in store for Reiser's filesystems.

Okay, a bit of history now. Before I start I want to mention that I may recollect a few things not quite correctly here as I read all this stuff quite some time ago. But I still think I shouldn't be too wrong about this.

History
ReiserFS seems to have caused quite a few heated discussions on the Linux Kernel Mailing List (Archive: http://www.lkml.org). Obviously enough that Hans Reiser sees the main reason for Reiser4 not being transferred into the main kernel-tree in people simply not liking him.
How much of this is true or not shall not be matter of this post, and anyway I am not in the position to judge about this. But it seems to be quite a fact that Reiser didn't make himself very popular with the code he submitted. The problem was, as far as I remember, that it was written without following any kernel-coding-standards.
Moreover it seems that more or less all work on ReiserFS (version 3) has stopped (except for simple bugfixes) when work on Reiser4 started, although there still were quite some complains about ReiserFS.

Reiser4 never made it into the standard kernel, probably due to different reasons, and thus is not available to the large mass of users who might get their hands on it otherwise. Patching the kernel yourself and rebuilding it might also not sound very encouraging for users to do just to try out a new filesystem.

And even if you decide to go that way, you'll still just have "some partition" as Reiser4, but not the root-partition, as that's needed during the process of patching and making.

Some more/other problems also are covered in the earlier linked Wikipedia-articles, so this should be it for this part.

Recent development
ReiserFS already seems to have suffered quite a lot under the process. Reiser's company Namesys is, contrary to early statements, either going down or has already come to an end.
Different pages (for example here on Wikipedia) you can find this comment made by a ReiserFS-developer and Namesys-employee:

Quote:

"Commercial activity of Namesys has stopped."

This, in my opinion, means it's over for Namesys. Although of course I could be wrong about this.

These two articles, A ReiserFS Without Hans Reiser and ReiserFS Without Hans Reiser, Continued suggest that work still goes on. But now these articles are 3 months old. So what has happened in the meantime?
It seems not much has actually happened during these 3 months. The latest files concerning the Reiser-filesystems (both version 3 and 4) now are located on kernel.org (ReiserFSProgs, Reiser4Progs and Reiser4 Kernel-Patches) but date back to April. So these files by now are at least 4 months old (I believe this is just the import-date to the server, but not the date these files actually have been created).

I guess that nobody in the kernel-community has the time (or is willing) to work on ReiserFS and the remaining Namesys-folks that continue their work (if anybody does) most probably focus on Reiser4.

Another blow is that Suse, formerly the biggest supporter of ReiserFS, also switched to ext3 a couple of versios ago. Thus all major distros now should use ext3, which of course does not help ReiserFS in terms of popularity.

And now?
What can we expect now? ReiserFS (3) doesn't seem to get much attention nowadays. The latest packages are pretty old already, which surely will not encourage distros and users to use the filesystem.
Reiser4 on the other hand has the problem that it still didn't make it into the standard-kernel and thus will most probably not find it's way to the common user.

Are ReiserFS and Reiser4 doomed to become extinct?
I think the answer is yes. Personally I think this is sad, ReiserFS, although surely not perfect, was and is an innovative FS.
Reiser4, judging from what I read about it, must be pretty interesting, also I have to admit that I didn't find the time to test it.

So I think the future will belong to ext4, especially since other interesting filesystems like XFS and JFS don't find a lot attention with users, most likely due to distros pretty much using offering ext3 as default choice. But still I would like to mention that XFS and JFS, especially XFS, really are interesting alternatives.
I once did some benchmarking (see here (German)) and XFS came out quite good.

I guess we will have to wait what will finally happen in the "case ReiserFS". As the saying goes: "Hope dies last." But I honestly think we shouldn't hope too much.

Comments

Well, I for one still use

Well, I for one still use ReiserFS on my biggest partition, /mnt/data, where I store stuff like movies, music, photos and other images, project files and some backups..

But that's mostly because I used it before and didn't reformat the partition in ages. Sticking out tongue If I were to create a new one I'd probably go with ext3, if nothing then because of a little less confidence about the continued support for ReiserFS, although if it's not being developed anymore then I suppose not much can go wrong so it's really a weak decision. The biggest threat I can imagine is ReiserFS being removed from the kernel, which probably wont happen any time soon, right?

Anyway, it's a good analysis and I wonder what others think. I dugg it. Smiling

Edit: Since our last poll is already quite dated this seemed like a good opportunity to add a new one: Are you abandoning ReiserFS?. Votes welcome. Smiling

Cheers

Thanks for digging it. I

 

Thanks for digging it. I wonder what the response will be.

I actually don't use ReiserFS anymore, and EasyLFS now throws an error if you want to use it as root-FS together with SELinux as it cannot handle the labels.
Otherwise I still support installing EasyLFS on ReiserFS, probably until it's removed from the kernel, and I have no idea if that will happen.

Digg comment.

I posted a comment on Digg about your blog entry. Excellent post BTW.

I have to admit that

 

I have to admit that reiserfs might eventually fall, and it's popularity has always been suffering, but I'm not ready yet to put the integrity of my files on another filesystem.

It has never failed me, so I believe I'll have it for a long time. Maybe until it starts getting unbearably buggy with new kernels, or removed from new kernels.

Reiser4 is unfortunately not in the kernel and I recently discovered this when I tried to build my system on a reiser4 formatted drive, and then found out that my kernel can't even boot it.

Simply stating it: it sucks. ReiserFS was, and still is for the time being, an incredibly stable filesystem, but if development stops, then I'm sure that'll go down the drain aswell.

 

The problem was, as far as I remember, that it was written without following any kernel-coding-standards.
Moreover it seems that more or less all work on ReiserFS (version 3) has stopped (except for simple bugfixes) when work on Reiser4 started, although there still were quite some complains about ReiserFS.

[link removed]

Re: Death of a filesystem(?)

[offtopic]

The comment above is spam, but instead of deleting it I only removed the link and kept it as a warning to spammers that this method doesn't work here either!

What method? This spammer copied a couple of sentences from the article itself to make it look like a perfectly valid and on-topic comment, but then linked to a completely offtopic site (about rapidshare links).

All subsequent spam comments using this method of spamming will be deleted without further warning.

[/offtopic]

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.