[NTLUG:Discuss] ext2 vs. ext3 vs. reiserfs

[Chris Cox]

Greg Edwards wrote:
...
> Couple different events caused the loss of the FS.  Once I turned on 
> gated with a wrong parameter and by the time I killed it the superblock 
> was toast.  When I ran fsck (reiserfs flavour) it asked about fixing 
> invalid nodes and like a dummy I said yes :(  After an hour I figured I 
> was in trouble, and I was right.  fsck couldn't figure out was FS was on 
> disk after that.

Hardly a FS problem.  When dd if=/dev/urandom of=/dev/hda .... what
are you going to do??  Can I kill my FS... absolutely.

...
> The only time I've lost an ext2 FS is after a hard crash.  I run UPSes 
> and rarely shutdown so ext2 is fine under normal use.  It'll be a long 
> time before I trust a critical system to a new FS again.
> 

I've had some problems with ext3, but mostly RH issues (things RH could
fix if they really wanted to... often associated with RH's lack of
knowledge about LVM).  I run reiser in production.  I have yet
to have a problem... but I use SuSE which is reiser/LVM friendly.  If you
are going to use ext3, RH is your friend... sometimes a stupid friend
however (can we get some old Unix guys to work for RH for awhile??).

There are times I wonder about ext2 as well... though I realize
it is considered to be the "rock" of Linux filesystems.  I'm not
sure of that though (but again, I could be seeing it through
RH tinted glasses).