xfsprogs

[Alma J Wetzker]

Collins Richey wrote:
> On Thu, 12 Aug 2004 13:19:04 -0500 (EST) Net Llama! <netllama at linux-sxs.org>
> wrote:
> 
>>On Thu, 12 Aug 2004, M.W. Chang wrote:
>>
>>>>What are you trying to fix?  I don't think journaling file systems need
>>>>fsck.
>>>
>>>I have once had ext3 corruption problem with my harddisk,
>>>demanding manual intervention.
>>>
> 
> 
>>fsck.xfs is a noop.  As others have noted, use xfs_repair instead, and
>>you'll be golden.  BTW, if xfs_repair runs and has to fix stuff it will
>>create a lost+found dir with the bits that had to fix.  You must delete
>>the contents of the lost+found dir before trying to run xfs_repair again,
>>or it will keep fixing the same thing over and over again.  This is by
>>design.
>>
> 
> 
> Thanks for that info. 
> 
> I echo the earlier comments: there is no such thing as a perfect fs, but some
> are less perfect than others.
> 
> I recently converted to xfs a few weeks ago after 3+ years on ext3 (no problems
> ever including after power failures). . Based on what I know about journaled fs,
> I would expect from time to time to have the occasional corrupted file after a
> less-than-orderly shutdown, but never the total fs losses (2x) I incurred with
> reiserfs. On the second occurrence, reiserfs managed to lose track of the
> primary copy of the  superblock and was unable to put the pieces back together
> from one of the copies! Never again in this lifetime.

FWIW, that matches my experience with reiserfs.  It may have gotten better 
since I used it, but it doesn't sound like it and I will never find out.  That 
is the type of experience that you just can't seem to let go of....

     -- Alma