Message NFS server not responding

[Ajay Kumar_5]

In syslog message is NFS server not responding. then NFS server ok. I am able to
go in to mounted directory .Pls help us what is the problem and how to rectify this .

Thanks Regards
Ajay

[Pete Randall]

Ajay,

It sounds like your system had a disk mounted from another system and that other system was temporarily unavailable (re-booted?) but is now back on line.


Pete

[Ajay Kumar_5]

The file system is NFS mounted in to the system where the message is coming. The message in syslog appeared no of times at irregular intervals . File system is ok and can access mounted file system.

Thanks and Regards
Ajay

[Jan Sladky]

hi Ajay,

for certainty check all demons needed on server side:

rcpbind (prior to hp-ux-10.20 called portmap)
nfsd
rpc.mountd
rpc.statd
rpc.lockd
rpc.pcnfsd - optional

br Jan

[sreejith_4]

Hi,

If you are using automounter , this problem can occure.
DO one thing..keep a shell opened from the remote mounted directory.
i,e. # cd /$nfs_ directory

watch whether it is giving the error after doing this.

Thanks
Sreejith M

[Muthukumar_5]

hai,

Your NFS server is rebooted or not-accessible for some time. That messages are stored in the syslog.

Use nfsstat command to get the nfs & rpc status. During that period,nfs mount directory files can not be accessed.

After the recovery,it is normal. If your mount or mount -v will be containing *not responding message for the mount's use umount to unmount the directory.

Check the server status nfsstat or reachability check with ping or linkloop. If it is succeed, then mount that directory.

Regards,
Muthukumar.

[Bharat Katkar]

In addition to all above check your Network as well...Could be a network congestion problem also.

Regards,

[Ajay Kumar_5]

Dear Muthukumar,
Thanks to all. I ahve checked all the suggetions given by all of you.In nfsstat i found in client rpc calls=38098 and badcalls=36 and client nfs calls=38062 and badcalls=0 . The output of nfsstat -m is as
/devtrans from dev:/usr/sap/trans Flags: vers=3,proto=tcp,auth=unix,hard,intr,link,symlink,devs,rsize=32768,wsize=32768,retrans=5
All: srtt= 0 ( 0ms), dev= 0 ( 0ms), cur= 0 ( 0ms).Also we are able to access that nfs mount file system. But when try to umount that the sollowing error is there
# umount /devtrans
nfs umount: nfs_unmount: /devtrans: is busy
But fuser shows no one is working in this.
Pls sugest. we have rebooted the server and again mounted the file system.But the message is coming in syslog.
Thanks and regards

[Muthukumar_5]

hai ajay,

I have reproduced your problem as,
another terminal user is within the mount directory.
terminal1>cd /dir_mnt

terminal2>umount /dir_mnt
nfs umount: nfs_unmount: /dir_mnt: is busy
umount: return error 1.

Make that no one within that mount directory. Else it will say this message

Regards,
Muthukumar.

[Anand_20]

Hi,

Probable culprit may be the lack of system resources- sometimes huge process take the memory pages and make the server unavailable.
Pl. monitor physical memory at regular intervals. and ofcourse
check the network I/O statistics.

This was one of my experience,

Anand

[Chris Saunderson]

Typically, this message points to resource constraint on the server involved (CPU, memory, IO, network) or possibly on the clients network.

HP-UX 11.00 had an "interesting" method of allowing NFS and buffer cache to interact. I had the pleasure of listening to Dave Olker describe the bad juju that happened and how tuning buffer cache can help immensely.