I am finishing a transactional storage system,
called "bedrock", to serve as a foundation of a
crash-proof database. I need to disallow NFS
access to the bedrock file, since transactional
locking can not be enforced across NFS. To this
end, I need to be able to answer, given a file
name, whether it resides on an NFS-mounted file
system, or locally. Actually, just the answer "is
it local?" will do.
Which Unix, actually Linux, system calls can be
used to determine the locality of a file?
Which system calls are used to allow root login
only from the physical console, for instance? Are
they different?
BTW, once the bedrock is finished, the technical
report is going to be published (in a week), and
the code will be GPLed on the web. One of the
uses is a
fsck-free Linux!
Basically, the bedrock can be mounted as a file
system, and then, after a crash, only the things
lost anyway will be absent -- no fsck ever again.
As I need the bedrock for application purposes,
I'm not going to have time I'd like to write a
Linux file system, but will be happy to assist a
serious guru to adapt the bedrock (easy) to make
one (hard, but now possible -- and zero cool).
--
Cheers,
Alexy Khrabrov -- www.suffix.com -- Segmentation f%^(&
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
