NFS readdir+ failed for server jserver: RPC: Timed out

[Bridget B]

Hi
Can anyone please help me out with this problem? I have one hpux server , exporting a filesystem to two different hpux workstations, all running HPUX11.
On workstation1, the nfs mount is ok and accessible by users who use it has thier home directory, for exmaple user fopitt. On workstation2, it mounts, but user fopitt ( same uids, and gids etc across both workstations), cannot access the mount.
In fact an "ls" as root in the directory where the NFS is mounted, returns the RPC error eventually and no directry listing. Both accounts use the nfs share as thier home directory.

The biod, rpc daemons etc are running ok on both workstations although workstation2 is running more biod than workstation1. Ping from workstation2 to the jserver is ok. nfstat -rc is ok, directory perms are ok.

I 've seen references to document #KBAN00000261 but i can't find it on the ITRC.

Any ideas/ suggestions/help would be welcome.
Thank you

[Peter Godron]

Hi Bridget,
and welcome to the forums !

The doc should be available at:
http://www4.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000079998580

Have you seen these earlier threads:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=103914http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1049487


Please also read:
http://forums1.itrc.hp.com/service/forums/helptips.do?#33 on how to reward any useful answers given to your questions.

[Steven E. Protter]

Shalom,

The file locking mechanism on NFS v3 differs from vxfs. This would explain some of the issues that you experience.

Other idea:

Does /etc/exports have any user/hostname filters on it? Quite possible.

SEP

[Bridget B]

Hi
Thank you for your welcome and suggestion but it looks like they refer to the same document/thread

nfsstat -rc gives:
Client rpc:
Connection oriented:
calls badcalls badxids
0 0 0
timeouts newcreds badverfs
0 0 0
timers cantconn nomem
0 0 0
interrupts
0
Connectionless oriented:
calls badcalls retrans
938 18 63
badxids timeouts waits
0 67 0
newcreds badverfs timers
0 0 54
toobig nomem cantsend
0 0 0
bufulocks 0


nsfstat -rs on the server ;
Server rpc:
Connection oriented:
calls badcalls nullrecv
0 0 0
badlen xdrcall dupchecks
0 0 0
dupreqs
0
Connectionless oriented:
calls badcalls nullrecv
531745 0 0
badlen xdrcall dupchecks
0 0 11914
dupreqs
0

The exports file on the server (jserver) reads
# cat /etc/exports
/WX

there are no options with the export line
mount line on workstation2, is the same as workstation1 :-
jserver:/WX/users/ /WX/users/ nfs rw,soft 0 0

Thank you

[Dave Olker]

Hello Bridget,

The fact that your clients are using UDP instead of TCP tells me you are likely running HP-UX 11.0 and either haven't patched these systems for a while or have never enabled the use of NFS/TCP. You might want to consider turning on NFS/TCP support and see if TCP allows your systems to work better.

If you're getting "RPC: Timed out" messages on an NFS/UDP mount, it usually indicates either a networking problem or the client/server is not sending/receiving packets in the allotted time.

Have you investigated your network to make sure there are no underlying problems (duplex mismatch, speed mismatch, port setting conflicts, etc.)? Have you checked your network switch to see if it is logging any errors? Do other network programs see the same problem (rcp, ftp, ttcp, netperf)? Are "netstat -s" or "lanadmin" logging any errors?

Again, if there is not an underlying problem in your network then my first suggestion would be to try using NFS/TCP and see if it behaves any better in your environment.

Regards,

Dave

[Bridget B]

thank you.
Now I know this is going to sound stupid ( its near the end of the day and my brain has gone to sleep), but can you please remind me how to switch it over to TCP NFS ?

[Dave Olker]

Sure. Install the latest NFS patches for 11.0 (just to make sure you have all the known fixes) and then issue the command:

# setoncenv NFS_TCP 1

This will enable NFS/TCP support.

# /sbin/init.d/nfs.client stop
# /sbin/init.d/nfs.server stop
# /sbin/init.d/nfs.server start
# /sbin/init.d/nfs.client start

This will stop/start the NFS client and server processes on the system. Alternately you can reboot the systems, but you don't have to.

You will need to do this on all of your HP-UX 11.0 systems - not just clients, but servers as well.

Regards,

Dave

[Bridget B]

Thank you all for your help- Unfortuantly this is pointing towards a network problem so I have left it with the network group to look into for now.