Re: Linux Distro

Nico wrote:
....
start failing, on any operating system. Everytime I've had a disk
start failing with ReiserFS, on either operating system, any attempts
to repair the apparent failures from the operating system corrupted
the ReiserFS and lost data, forcing me to restore from backup after I
replaced hardware.
It is very true that reiserfs doesn't work well on drives with
failing sectors. But this is a limitation with many filesystems.
In fact, filesystems that mask the problem are only hiding
a problem that needs to be corrected.

"Doesn't work well" misses the point. It wreaks random file system
destruction throughout the file system. Zeroing files randomly is
*not* acceptable behavior.

It's not exactly random behavior, it's more of a by product
of the layout it uses on the disk. You could say that because
it's a more complex data structure on the disk, that it is
more susceptible to hardware issues. That would be more
accurate. You're attempting to paint the picture darker
than what it truly is. But you have certainly made your
point... we know where you stand with regards to your
likes and dislikes. I'm merely trying to correct the few
point of (sometimes) extreme misinformation you are giving
along side your bias.


Again, if using Red Hat, it could well be that corruptions
will abound. But not so with SUSE. The problems you are
describing only happen when something bad has already
happened. And... again, ext3 isn't immune... I've seen
plenty of ext3 corruptions that ext3's fsck couldn't fix.
But again, there's not much you can do when the hardware
is at fault.

Wrong. Both operating sytems, different classes of hardware failures.
(The RHEL system had the IBM deathstar drives, the SuSE system
had..... other issues.)

What I mean is that since SUSE used reiserfs since 6.3 and as
it's primary filesystem since 6.4, that they've had a vested
interest in its success/stability. The SUSE team contributed
much to the filesystem over the years and have fixed numerous
problems and added many features. Since Red Hat didn't care
so much, their addition of critical patches was too slow and
out of sync with the more stable release points of reiserfs,
leaving some interesting bugs inside of reiserfs in many of
the Red Hat releases... but again, like you, Red Hat has
their bias, it simply wasn't that interesting for them to fix.


"Can't fix" is one thing. "Corrupting much of the file system" is
another story. Fortunately, I'm a paranoid peson and insist on regular
backups, especially when a system begins to fail. But for one system
(the SuSE), it could no longer be gracefully backed up before
proceeding. And the backup failures were the sign of trouble
occurring.

I understand your frustration. I'm just pointing out that while
ext3/4 might SEEM less prone to errors when dealing with failing
drive data, it's really not so true. It's just that the problems
will stand out more readily with reiserfs... that's all. The problem
is still there AND it can lead to interesting problems for ext3/4
as well.


File system tools shouldn't do that. But then, its author Hans Reiser
seems to have a real gift for losing important things and being unable
to explain what happened to them. (Despite the missing body's blood
stains in his car and being found with his passport and $9000 in cash:
no, I'm not kidding, he's on trial for murdering his ex-wife.)
This also somewhat true (though obviously biased). Hans could be in for
some real trouble, but obviously he hasn't be convicted of anything.

Personally, I pray that he didn't do it. And if he did, I hope that
he'll feel remorse and pay for his crime.

His guilt seems.... seriously likely. As harsh as it seems, the
details of his behavior, I'm afraid, seem familiar to me from trying
to revoer from a ReiserFS failure. The evidence is concealed only from
casual perusal, but digging into and observing other logs more
carefully shows damning traces of whose fault it was, even if every
detail is not clear.


Hans seems to be a very interesting/disturbed fellow. However, he
also seems to possess a genius quality. I think his technical
abilities are very good. But, like you, I fear that the evidence
so far doesn't paint a pretty picture for Hans' future. We'll see.


.

Relevant Pages

  • Re: [SLE] frequent hard disk failures - 9.3?
    ... Carlos F Lange wrote: ... > disks either crashing or showing bad sectors. ... SuSE and ReiserFS were the only things in common ... so guess under what conditions failures occur. ...
    (SuSE)
  • Re: [SLE] frequent hard disk failures - 9.3?
    ... On Thursday 03 November 2005 21:42, Sid Boyce wrote: ... so guess under what conditions failures occur. ... I shot down reiserfs as the cause the ... on the hard drives, ...
    (SuSE)
  • Re: Too much hard drives failing
    ... hardware or disk corruption/software? ... using Reiser 3, I also had failures which I eventually traced to 3ware controller firmware, however I have not had any hardware failures. ... Can this anyhow be related with reiserfs? ...
    (Fedora)
  • Re: [opensuse] RE: Siutable File Systems
    ... When it fails--- its not reiserfs ... What is interesting is that an HD will run with failures on EXT3 where the ... drive Reiser4 towards mainline inclusion in the kernel. ... might be better to say you lost a drive vs you lost data. ...
    (SuSE)
  • A reiserfs catastrophe?
    ... One of my Suse 9.1 boxes failed a couple days ago. ... 40GB reiserfs partition for /, and a second reiserfs partition of 35GB ... perfect sense, for a failed drive/filesystem, but not a Linux crash. ...
    (comp.os.linux.misc)