Dale Amon wrote:
> On Tue, Jun 18, 2002 at 09:02:32PM +0300, Jari Ruusu wrote:
> > Problem is that reiserfs guarantees only meta-data integrity but does not
> > guarantee file-data integrity in system crash situation, yet it assumes that
> > all lower layer devices do provide data integrity and ordered writes. When
> > you put journaled fs on loop that can't provide data integrity and ordered
> > writes all the way down to disk platters, you have problems.
> >
>
> Ah, so you've now explained the reason why I had the only fault I've
> ever seen on cryptoapi. I had this on my laptop:
>
> > reiserfs -> loop -> reiserfs -> partition
#2 #1
Here reiserfs #1 does not guarantee ordered writes, so fs corruption is
possible on reiserfs #2.
> and I've had problems with corruption on occasion. However I've had
>
> > reiserfs -> loop -> raid1 -> partitions
>
> and have had about 2 years of trouble free operation.
Here loop and raid1 provide ordered writes so no corruption.
Regards,
Jari Ruusu <jari.ruusu@pp.inet.fi>
-
Linux-crypto: cryptography in and on the Linux system
Archive: http://mail.nl.linux.org/linux-crypto/
