Re: Diskless RH6.0 problem

[pmd <pmd@azad.demon.co.uk>]

[...]
> Now I have another problem:
> When I do /etc/rc.d/init/atd stop
> I get the message
> Stopping at daemon: NFS: can't silly-delete run/.nfs001a38580000000a, error=-13
> and the 0 byte file /var/run/.nfs001a38580000000a suddenly exists.
> Every time I do that I get a new 0 byte file with almost the same name.
> 
> I have isolated the problem a little. If I do
F
> /etc/rc.d/init/atd start
> and then killall -TERM atd 
> I get the same problem. Killall -KILL atd don't! 
> 
> Some ideas?

This is mostly a guess but looking at the code fs/nfs/dir.c it seems
that when a process deletes a file that is open by some process the
local NFS system has to rename it to a 'hidden' .nfs file (to preserve
normal Unix semantics) and when the use count for the file goes to zero,
because it has finally been closed, it deletes it.

I imagine something has /var/run/atd (where the PID of atd was stored)
open when killproc is called from init.d/atd which removes the file so
the rename happens but by the time that process exits the permissions
have changed or maybe /var is no longer mounted.

Does something remount /var during startup after /var/run/xxx has been
created with different permissions?

--
Phil Davis  - pmd@azad.demon.co.uk

===========================================================================
This Mail was sent to netboot mailing list by:
pmd <pmd@azad.demon.co.uk>
To get help about this list, send a mail with 'help' as the only string in
it's body to majordomo@baghira.han.de. If you have problems with this list,
send a mail to netboot-owner@baghira.han.de.