Richard A Downing wrote:
I think that your solution, while not seemingly elegant, is a good
workaround. rpc.nfsd is just a userland program to tell the kernel
module how many threads to run. As such, it reports success if it told
the kernel OK - what the kernel does with this info is a different
matter. Your solution correctly tests 'what the kernel did'. So I
think it's right.
Thanks. :-)
The only real question left is 'Why do you want to know?'. Do you abort
the shurdown if the nfsd threads don't die, or carry on with a worry
that some data wasn't safely written? Not sure what the right answer is
here.
Well, thanks again, didn't think about that. I've expanded the command a
little bit. It's not looking good, but should hopefully provide a little
safety:
if pidof nfsd 2>&1 > /dev/null ; then {
rpc.nfsd -- 0
sleep 1
if pidof nfsd 2>&1 > /dev/null ; then {
echo_failure
boot_mesg "Killing NFS nfsd..."
kill -HUP 'pidofproc nfsd'
evaluate_retval
sleep 10
}
else echo_ok
fi
}
else echo_warning
fi
Haven't had chance to test this yet, and guess I have to fake a nfsd
lockup to see what happens in worst case.
My attempt is based upon Richard Downing and Bruce Dubbs tips from the
BLFS-dev list. Couldn't find a good solution there though, which is why
I'm trying this on my own.
Again, all comment, corrections or better solutions are very welcome.
--Tor Olav
--
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
