Re: Open by inode? (was Re: knfsd)

david parsons (o.r.c@p.e.l.l.p.o.r.t.l.a.n.d.o.r.u.s)
5 Jan 1999 00:40:14 -0800


In article <linux.kernel.19990103162906.C4142@tantalophile.demon.co.uk>,
Jamie Lokier <lkd@tantalophile.demon.co.uk> wrote:
>On Sat, Jan 02, 1999 at 12:29:48PM -0500, Chip Salzenberg wrote:
>> According to Alan Cox:
>> > NFS likes to work at the inode level and its awkward to get at that
>> > [from user space]. If you could open by inode with some other stuff
>> > from the kernel hooked and write a clone based UNFSD then probably
>> > [knfsd would be unnecessary].
>>
>> Is there any reason why the superuser shouldn't be able to open by
>> inode? This seems an obvious idea; perhaps there's also an obvious
>> argument against it that I'm not seeing...
>
>Is this really necessary?

From my experience, I'd say yes; I've a largish (26gb) NFS file
server that's being run at a company I sysadmin at, and it has the
annoying problem of generating hashes that are almost but not quite
unique -- a (mv;mkdir;mv;ls) sequence will pretty regularly (once
every two-three days, if the bug reports are to be believed) lead
to either the moved file or created directory being invisible until
the nfsd cache drains.

____
david parsons \bi/ A "why can't commercial Unices use smb?" moment if
\/ there ever was one.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/