'nfs.server stop' does not work on Itanium 11.23

[Bob_Vance]

HPUX 11.23
rx1600 itanium

*********************************
NFS server starts up fine at boot.
*********************************

It also will start if I
. disable NFS server (NFS_SERVER=0 in nfsconf)
. reboot
. re-enable NFS (NFS_SERVER=1 in nfsconf)
. 'nfs.server start'

I get processes like this:

daytona2 ## psg nfs
... root 9 0 0 08:48:52 ? 0:00 nfsktcpd
... root 1708 1702 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1707 1701 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1702 1701 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1700 1 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1701 1 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1703 1701 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1704 1702 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1710 1701 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1711 1701 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1712 1701 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1713 1701 0 08:51:15 ? 0:00 /usr/sbin/nfsd 16
... root 1714 1702 0 08:51:16 ? 0:00 /usr/sbin/nfsd 16
... root 1716 1702 0 08:51:16 ? 0:00 /usr/sbin/nfsd 16
... root 1717 1701 0 08:51:16 ? 0:00 /usr/sbin/nfsd 16
... root 1718 1702 0 08:51:16 ? 0:00 /usr/sbin/nfsd 16
... root 1719 1702 0 08:51:16 ? 0:00 /usr/sbin/nfsd 16
... root 1720 1702 0 08:51:16 ? 0:00 /usr/sbin/nfsd 16

And it works -- clients can mount stuff.


**********
However,
**********
when I try a
... /sbin/init.d/nfs.server stop
not all the 'nfsd' processes get killed.

Furthermore, I cannot kill the remaining ones, even with a 'kill -9 ...'

No clients have mounted anything, so it is clean -- just after a reboot.

At this point, even if I manually start
'rcp.mountd'
clients still cannot mount anything.

A RESTART doesn't help either:

daytona2 ## /init.d/nfs.server start
.. starting NFS SERVER networking
..... starting up the rpcbind daemon
........ rpcbind already started, using pid: 796
..... Reading in /etc/exports
..... starting up the mount daemon
........ rpc.mountd already started, using pid: 14631
..... starting up the NFS daemons
........ nfsd(s) already started, using pid(s): 1707 1702 1701 1712 ...
..... starting up the Status Monitor daemon
........ rpc.statd already started, using pid: 842
..... starting up the Lock Manager daemon
........ rpc.lockd already started, using pid: 848

clients still cannot mount anything.


******************************
I have the latest NFS patches:
******************************

daytona2 ## grep -i nfs /tmp/swlistpr
... PHCO_32523 1.0 quota(1) on an NFS client;uid enhancement
... PHKL_32713 1.0 nfs client hang; cache FS umount
... PHNE_33098 1.0 Core NFS cumulative patch
... PHNE_33655 1.0 NFS cumulative patch

Actually, PHNE_33655 has been superseded by PHNE_33982,
but PHNE_33982 does not help, either.


This procedure works perfectly fine on 11.11, of course,
but I do not have an 11.23 PA-RISC system to check.

Has anyone else seen this on Itanium 11.23??
Or do they have one where it works fine ?

tks
bv

[Geoff Wild]

Before attempting to stop the server, you should stop the client first.

Correct order to recycle NFS is:

/sbin/init.d/nfs.client stop
/sbin/init.d/nfs.server stop
/sbin/init.d/nfs.core stop


/sbin/init.d/nfs.core start
/sbin/init.d/nfs.server start
/sbin/init.d/nfs.client start


Rgds...Geoff

[Bob_Vance]

Thanks, Geoff

But I only want to stop the server.
You can certainly stop the server without stopping the client -- there is logic in the scripts to handle this and it seems to work fine on 11.11.

Further, the problem (11.23 NFS server unable to be stopped) exists even when client has never been started; e.g., NFS server-only system.

Anyway, stopping client first made no difference to the problem on Itanium 11.23.


tks
bv

[Geoff Wild · Accepted solution]

Okay - I'll try it - I have 1 Itanium box running 11.23:

root@sha1 [ / ]
# ps -ef |grep nfs
root 10 0 0 Jun 20 ? 0:02 nfsktcpd
root 3363 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3345 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3367 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3359 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3365 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3338 1 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3347 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3340 1 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3341 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3351 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3356 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3353 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3370 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3372 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3373 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3374 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16
root 3375 3340 0 Jun 20 ? 0:00 /usr/sbin/nfsd 16

root@sha1 [ / ]
# /sbin/init.d/nfs.server stop
killing nfsd
killing rpc.mountd

root@sha1 [ / ]
# ps -ef |grep nfs
root 10 0 0 Jun 20 ? 0:02 nfsktcpd



And startup worked:

root@sha1 [ / ]
# /sbin/init.d/nfs.server start
starting NFS SERVER networking

starting up the rpcbind daemon
rpcbind already started, using pid: 1313
starting up the mount daemon
/usr/sbin/rpc.mountd
starting up the NFS daemons
/usr/sbin/nfsd 16
starting up the Status Monitor daemon
rpc.statd already started, using pid: 1434
starting up the Lock Manager daemon
rpc.lockd already started, using pid: 1440

Here's the version I'm running:

# swlist |grep OE
HPUX11i-OE-MC B.11.23.0512 HP-UX Mission Critical Operating Environment Component


Here's my patches:

# show_patches |grep -i nfs
PHCO_31645 Japanese NFS/LIBNSL manpages
PHCO_31647 Japanese NFS manpages
PHCO_31661 Japanese NFS/RCMDS manpages
PHCO_31662 Japanese NFS/NIS manpages
PHCO_31666 Japanese core NFS manpages
PHCO_32523 quota(1) on an NFS client;uid enhancement
PHKL_32713 nfs client hang; cache FS umount
PHNE_32055 NFS cumulative patch
PHNE_33098 Core NFS cumulative patch

Rgds...Geoff

[Bob_Vance]

Thanks, Geoff.

That proves that it is supposed to work :>)

Now, the question is why not on my systems :[

The only patch difference I see here is that I have tried 2 later NFS cumulative patches than your PHNE_32055 (PHNE_33655 and PHNE_33982).

However, my OE is also a little earlier (sep04)

HPUX11i-OE-MC B.11.23.0409 HP-UX Mission Critical Operating Environment Component

I guess I'm just stuck.

bv

[Geoff Wild]

Yeah - I loaded the Dec05 version in January. So far - it's been very stable. This is on my workstation - I usually run new releases on it for 6 months prior to moving to any test servers, then I roll out to production.

As far as your issue - I would put a cal into HP - as what you want to do does work - just not on that server for some reason.

Rgds...Geoff

[Bob_Vance]

Agreed.
Thanks, Geoff.

bv

[Bob_Vance]

Yesterday, I updated the system with all the latest patches.

They included a new NFS cum patch.
This seems to have fixed the problem:

old patch
... PHNE_33982 1.0 NFS cumulative patch

new patch
... PHNE_35118 1.0 NFS cumulative patch

tks
bv