[sclug] Well, *that's* not supposed to happen...

David Given dg at cowlark.com
Thu May 24 23:35:44 UTC 2007

Hash: SHA1

Andy Hayward wrote:
> Reiserfs is awful, and I wouldn't trust it with any data that I
> actually cared about.

Yes, I remember you telling me about it --- that's one of the reasons why I
don't use it any more!

> There are known attacks against all three hash functions used for the
> directory index.

I'm not actually entirely convinced. Firstly, hash collisions don't appear to
actually cause data *loss*, they simply (from my understanding) result in the
inability to create a file when such a collision would occur --- the
filesystem remains consistent. (Of course, if something doesn't handle the
error condition properly and propagates the error, that might cause further
problems.) Such an error should in theory be no more fatal than, say, running
out of disk space.

Secondly, the messages you quote do come from 2004, and I can't find any more
recent reference to the issue. I would have thought that such a serious
problem would have raised a bit more of a fuss, but distros like Debian and
Ubuntu still have it as the default option. It still happens --- I've just
tested it, with the script supplied --- but it's apparently, rightly or
wrongly, not considered that serious.

The main reason I don't use reiserfs, however, isn't anything to do with that
bug; it's because of the endless bickering and arguments between the reiserfs
team and the Linux kernel developers. It betrays a lack of professionalism
that's... disturbing. The insistence on violating the filesystem API for
ReiserFS 4, for example. The reiserfs people may think it's necessary, but if
the kernel maintainers decide that it's a bad idea, that should stop doing it,

*shrug* JFS works fine for me.

> Both JFS and XFS have their strengths, but for the vast majority of
> purposes ext3 (with dir_index enabled and data=journal) is sufficient.

Well, as I said, I've never been able to make it work. I know it works for
other people, but frankly I'm not interested enough to put the research in to
find out *why* it didn't work. There are other filesystems that are more
efficient and faster and easier to set up.

- --
??? ?????????????? ??? http://www.cowlark.com ???????????????????
? "Parents let children ride bicycles on the street. But parents do not
? allow children to hear vulgar words. Therefore we can deduce that cursing
? is more dangerous than being hit by a car." --- Scott Adams
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


More information about the Sclug mailing list