Re: NFS Problem in Kernel 2.0.27: inode status not updated

Top Page
Delete this message
Reply to this message
Author: Christoph Lameter
Date:  
To: Piete Brooks
CC: exim-users
Subject: Re: NFS Problem in Kernel 2.0.27: inode status not updated
On Tue, 31 Dec 1996, Piete Brooks wrote:

Piete.Brooks >> MUAs use lock file based locking under Linux
Piete.Brooks >
Piete.Brooks >Not all ..... :-(
Piete.Brooks >
Piete.Brooks >> which is the standard because of the NFS troubles.
Piete.Brooks >
Piete.Brooks >Err -- forgive me if I'm wrong, but the impression I get is that the algorithm
Piete.Brooks >based on hitching posts which at least some MUA/MTAs have been using under
Piete.Brooks >Linux do not work with any existing kernels ...

Strangely they for years worked for us. Only exim failed.
Other MUAs might check the return code of the system calls instead.

POSIX NFS standards do not guarantee the accuracy of the inode
information. Exim is definitely going beyond POSIX specs here which is
bad and I would consider Exim buggy not Linux.

--- +++ --- +++ --- +++ --- +++ --- +++ --- +++ --- +++ ---
PGP Public Key = FB 9B 31 21 04 1E 3A 33 C7 62 2F C0 CD 81 CA B5