Stales NFS file handle

[Dewa Negara_1]

I got an error message "Stale NFS file handle" when run bdf. I can not unmount the NFs file system because of busy. How can I unmount and re-mount the NFS file system? Any patch must be installed to solve the problem? Anybody could help me on this. Many thanks.

Rgds,
dewa

[root@bdhp6010:/var/adm/syslog]
# bdf | grep N5A
bdf: /sapmnt/N5A: Stale NFS file handle
2097152 691412 1317910 34% /usr/sap/N5A

[root@bdhp6010:/var/adm/syslog]
# umount /sapmnt/N5A
nfs umount: nfs_unmount: /sapmnt/N5A: is busy

[root@bdhp6010:/var/adm/syslog]
# mount -v | grep -i /sapmnt/N5A
bdhp4136:/sapmnt/N5A on /sapmnt/N5A type nfs defaults,NFSv3 on Thu May 30 17:56:
07 2002

[root@bdhp6010:/var/adm/syslog]
# mount -v | grep -i /sapmnt/N5A
bdhp4136:/sapmnt/N5A on /sapmnt/N5A type nfs defaults,NFSv3 on Thu May 30 17:56:07 2002

[S.K. Chan]

Check this thread for the troubleshooting tips and possible scenario of having different NFS version (reply by Paula)..
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x84d95a3e6e0bd6118ff40090279cd0f9,00.html

[Mike Hassell]

Dewa,

Check to see what has it's hooks into your NFS mount:

fuser bdhp4136:/sapmnt/N5A

If you see that processes are trying to access that NFS mount, then kill them with fuser:

fuser -k bdhp4136:/sapmnt/N5A

then try unmounting it:

umount bdhp4136:/sapmnt/N5A

If that doesn't work, you may have to reboot. All the above and more is included in S.K. Chan's above link. I would also recommend that you search on ITRC for the latest NFS v3 patches for your OS level.

Hope that helps.

-Mike

[Mihaita Bighiu]

Hi,

The best way is to reboot the system,but you could avoit the rrboot doing:

# fuser -kuc /usr/sap/N5A
# fuser -kuc /sapmnt/N5A
# umount /sapmnt/N5A
# umount /usr/sap/N5A

If this one does not work you have to stop/start MFS ( server client & core).
#/sbin/int.d/nfs.server stop
#/sbin/int.d/nfs.client stop
#/sbin/int.d/nfs.core stop
#/sbin/int.d/nfs.core start
#/sbin/int.d/nfs.client start
#/sbin/int.d/nfs.server start

This worked for me. If you are using automounter you have alto to restart the automounter.

If you have downtime you could do from console or web-console:
# init 1
# init 3
This will stop/restart all the network and NFS daemons.

MB