Re: [cobalt-users] [Raq2/Qube2] Sleeping nfsoid loaded in Kernels?

["James Hoaggs" <james_hoaggs@xxxxxxxxxx>]

We were able to do a "kill -9 6" and for processes 7, 8, and 9 and now
have a bunch of zombies by doing the "top" command:

 root            6  0.0  0.0  0   0  ?  Z Nov 10 0:00 (nfsiod (zombie))
 root            7  0.0  0.0  0   0  ?  Z Nov 10 0:00 (nfsiod (zombie))
 root            8  0.0  0.0  0   0  ?  Z Nov 10 0:00 (nfsiod (zombie))
 root            9  0.0  0.0  0   0  ?  Z Nov 10 0:00 (nfsiod (zombie))

These nfsiod zombies come back to life on re-boot and continue to exist
as sleeping and swapped out (SW).  Thinking that nfs is compiled in the
kernel, we checked out by doing vi /etc/rc.sysinit and found:
# Mount all other filesystems (except for NFS)....
mount -a -t nonfs

We would still like to permamentally remove these nfsoids, since we will
never want to mount our filesytems via NFS. We are going to do some tests
and see if we can remove the NFS subsystem from the kernel modules, unless
there are other alternatives.

> :>Our machines for some reason run client processes that are stuck
> in the
> :>kernel for a while as "nfsoid".  NFS is a connectionless and stateless
> :>protocol where NFS servers run on unpredictable port numbers via
> UDP
> :>or TCP.  Is NFS still sleeping somewhere within the NFS code via
> the
> :>kernel or modules based on our results below?
> 
> [snip]
> 
> :>So we did:
> :>[root@www /]# /usr/sbin/rpc.nfsd --version   or
> :>[root@www /]# /usr/sbin/showmount --version
> :>we received:
> :>Universal NFS Server 2.2beta37
> :>
> :>When we do a /etc/rc.d/init.d/nfs status we get:
> :>rpc.mountd is stopped
> :>rpc.nfsd is stopped
> 
> :>By issuing the command it looks like it just stopped,
> :>just like when you stop ssh and such.
> :>NFS is very insecure - file systems can be mounted and remounted
> even
> :>if the computer is re-booted, shutdown, or even disk copied to an
> other
> :>hard drive, the file systems can still be accessed at a later time.
> Hence
> :>we ask how to remove nfsoid from being generated at startup.
> 
> Getting exactly the same result. Even though asleep and swapped out
> (SW), I
> wonder what it would take to activate them?
-- 
James Hoaggs   ICQ #96365505
james_hoaggs@xxxxxxxxxx - email
(408) 380-2271 x8024 - voicemail/fax

__________________________________________________
FREE voicemail, email, and fax...all in one place.
Sign Up Now! http://www.onebox.com