2001-10-14 18:20:18

by Jens Benecke

[permalink] [raw]
Subject: Re: [reiserfs-list] Re: ReiserFS data corruption in very simple configuration

On Sun, Oct 14, 2001 at 10:52:54AM -0400, Chris Mason wrote:

> > When I was using ext2 I always mounted the /usr partition read-only, so
> > that a fsck weren't necessary at boot - and the files were all
> > guaranteed to be OK to bring the system up at least.
> >
> > Does this (mount -o ro) make sense with ReiserFS as well? What I mean
> > is, is there a chance of a file getting corrupted that was only *read*
> > (not *written*) at or before a power outage?
>
> Yes, after the mount is finished, reiserfs won't change the files on a
> readonly mount.

What I meant is this: AFAIK, if you exclude broken hardware, in ext2 there
is no chance of a file that was never written to since mounting being
corrupted on a crash, even if the fs was mounted read-write.

Is this the same thing with ReiserFS?


--
Jens Benecke ???????? http://www.hitchhikers.de/ - Europas Mitfahrzentrale

Crypto regulations will only hinder criminals who obey the law.


Attachments:
(No filename) (952.00 B)
(No filename) (240.00 B)
Download all attachments

2001-10-14 23:32:22

by Bernd Eckenfels

[permalink] [raw]
Subject: Re: [reiserfs-list] Re: ReiserFS data corruption in very simple configuration

In article <[email protected]> you wrote:
> What I meant is this: AFAIK, if you exclude broken hardware, in ext2 there
> is no chance of a file that was never written to since mounting being
> corrupted on a crash

Well, you can eighter lose the file due to a broken directory (maybe you
find the missing inode in lost+found) or it can even corrupt the file due to
a ext2 software error, which is unlikely but all filesystems in development
are reported to eat files every now and then.

Greetings
Bernd