Also Sprach Anthony A. D. Talltree:

> > http://lwn.net/2001/0503/a/lt-dump.php3
> >for what Linus has to say about it.
>
> Which as always is handwaving to cover the fact that he's too lazy to fix it.
>

I'm still confused about this. Is it because the kernel buffer vs page
cache architecture is inherently flawed, implemented poorly, or is
ext2/3 dump the problem? Since xfsdump and XFS are not subject to
the same caveats, it would seem possible to write a dump util for
a filesystem in the 2.4.x kernels which works subject to the usual
limitations of dump on a live filesystem.

Linus does mention in the above:

However, it may be that in the long run it would be advantageous to have a
"filesystem maintenance interface" for doing things like backups and
defragmentation..

Is such an interface necessary? I doubt FAT16/32 has such an interface,
perhaps it is so simple it doesn't require one. Syncsoft's backup util
for Windows NT/2000 claims it can backup live filesystems without
file corruption - does NTFS have such an API?

-- 
C. Chan < [EMAIL PROTECTED] >
PGP Public Key: finger [EMAIL PROTECTED]

Reply via email to