3

I've seen lots of comments suggesting that EXT3 is safer & better than reiserfs. I've been using this FS for about 4 years now and never ever had problems.

Before this I used ext2 and then ext3. On ext2 I once lost ALL my files because of a hardware problem, on ext3 a big part of my FS got rearranged and moved to "lost'n'found". On reiser I never ever lost a file, so I don't really understand why people keep suggesting to use ext3 as long as reiser does its job and seems to do it better.

The problems I had were caused by the IDE cable "disconnecting" from time to time from the HDD interface... while the system was running. It was a very specific situation when this happened and it took a really really long time until I changed the cable with a newer one...

I understand that Hans Reiser is in prison now... but why would that matter? We used FAT16/32 for a really really long time now and nobody complained. Nobody even knows if the original developer for that FS is still alive...

squillman
  • 37,618
  • 10
  • 90
  • 145
quamis
  • 362
  • 7
  • 18
  • "Nobody even knows if the original developer for [FAT] is still alive." Actually, we do. You might have heard of him: Bill Gates. – wfaulk Sep 26 '09 at 02:43
  • @wfaulk The credited author of FAT is [Marc McDonald](https://en.wikipedia.org/wiki/Marc_McDonald), Microsoft Employee #1. – Michael Hampton Aug 25 '12 at 20:47

4 Answers4

14

Its performance doesn't scale well with multiple processing cores due to the specific implementation.

Prior versions were actually fairly prone to corruption (and I've encountered this multiple times), though that's supposedly resolved in current versions.

Its fsck can actually end up causing more corruption on an already heavily corrupted filesystem.

Upgrading to its successor, Reiser4, requires a complete backup and restore, since it's effectively a completely different filesystem, unlike ext2->ext3->ext4.

There are very few developers actually working on Reiserfs, which means that if and when problems occur, there's not much support to fix them.

Much of this information is from one of the head SuSE developers when they decided to switch their default filesystem from Reiserfs to ext3.

wfaulk
  • 6,828
  • 7
  • 45
  • 75
  • I didn't know about the multiple processor problem, but the fsck works ok for me... etx3 failed me, reiser didn't. – quamis Sep 25 '09 at 16:43
  • 7
    While it's too bad that you lost data with ext3, no filesystem is going to be able to recover from all errors. If your point is "I never lost data with reiserfs, therefore it must not be as bad as people say", that's what is known as "anecdotal evidence" and isn't exactly a valid argument. It's rather like saying "I've never died from driving 150mph, therefore it must not be as dangerous as people say". – wfaulk Sep 25 '09 at 17:18
2

There are non-technical reasons reiserfs is marginalized. It is not actively being supported as much as it used to be, especially with the founder leaving the project.

TomOnTime
  • 7,567
  • 6
  • 28
  • 51
1

With his conviction, there seems to have been no one to hold Namesys together, and previously, they had real problems trying to get Reiser4 in to the kernel, for a mix of political and practical reasons (your bits need to fit the kernel coding style, for instance, and changing other things in the kernel is hard to get people to agree to).

It's a pity, as Hans was one of the few people who was trying for more than a better implementation of a POSIX file system, but it seems that Btrfs is taking up that torch.

ZFS is the other cool thing out there, but not for Linux.

Ronald Pottol
  • 1,683
  • 1
  • 11
  • 19
0

I've got machines on both reiserfs and ext3 and from a totally non-scientific observation of my recent personal usage, ext3 seems to suffer more than reiserfs. With reiser, I never had to rescue an un-bootable system whereas this has happened several times on ext3. I'd like to use ext3 for Linux but it borked out too many times for me.

sybreon
  • 7,357
  • 1
  • 19
  • 19