[SLUG] was: Sarasota Meeting Update

From: Bryan J. Smith (b.j.smith@ieee.org)
Date: Sat Sep 11 2004 - 08:56:15 EDT


On Sat, 2004-09-11 at 05:04, Pete Theisen wrote:
> Rather than silently leave corrupted files on disk,

Unless you force a reboot to do that (which is an option), or manually
force a mount without a consistency check (by using tune2fs to flip the
bit), this is virtually impossible with Linux.

It's the realm of FAT filesystems and the OSes that use them.

> or require an extensive "fsck" or "chkdsk" on recovery from a crash,
> a journaled file system like ext3 or ReiserFS can re-run or rollback
> updates to the file system to put it into a predictable and consistent
> state.

Typically. I've found both Ext3 and ReiserFS are very conservative and
respected in this regard. Unfortunately, because the internal structure
of ReiserFS changes by design, and I have caught at least one
distributor *COUGH*Mandrake*COUGH* not bothering to make sure the kernel
and user-space recovery tools match, I don't trust ReiserFS. If
ReiserFS drops out and requires an off-line repair, I immediately back
it up "raw" (dd) _before_ attempting to run those tools. This is in
addition to researching the _exact_ ReiserFS version in the kernel and
verify the user-space tools are 100% compatible.

[ NOTE: SuSE does probably the best job with ReiserFS. Better yet,
even one of their developers was honest enough to tell me to stick with
Ext3 back in 2000-2001 because of my requirements. ]

I also trust XFS explicitly. Other than one bug with XFS 1.1 that was
caught and fixed (it toasted the /var filesystem on two of my systems),
XFS is of the same proven quality as Ext3. It's internal structure
hasn't changed since it's introduction in the mid-'90s, and they ported
_everything_ over from IRIX almost verbatim. The 2.6 kernel's VFS also
benefits from a lot of SGI's GPL donations.

Of course, many 2.4 distributions did _not_ ship a completely XFS
kernel/user-space implementation *COUGH*Mandrake*COUGH*.

All my knowledge of JFS is second-hand (no personal experience). Since
it was ported from OS/2 and not AIX and missing a lot of UNIX compatible
interfaces that had to be retrofitted "clean room" (unlike XFS), for
what reasons I was dumb-founded until I read up on Project Monterey
later (i.e., SCO), it was the least compatible of any filesystem for my
needs when I first looked at it.

-- 
Bryan J. Smith                                  b.j.smith@ieee.org 
------------------------------------------------------------------ 
"Communities don't have rights. Only individuals in the community
 have rights. ... That idea of community rights is firmly rooted
 in the 'Communist Manifesto.'" -- Michael Badnarik

----------------------------------------------------------------------- This list is provided as an unmoderated internet service by Networked Knowledge Systems (NKS). Views and opinions expressed in messages posted are those of the author and do not necessarily reflect the official policy or position of NKS or any of its employees.



This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 17:21:23 EDT