NFS connection time out on Jaunty, working on Hardy

Asked by Martin Frauenstein on 2009-07-14

We have 2 servers, 1 IBM iSeries with 3 exported NFS shares another HP with Hardy server and 1 NFS exported share. There are 4 laptops in the local network, 3 with Hardy and 1 recently installed with Jaunty. The Jaunty installation does not connect to the IBM iSeries NFS shares. Below is an extract from the kern.log
Jul 14 01:49:02 Ferrari kernel: [ 83.592055] eth0: no IPv6 routers present
Jul 14 01:49:05 Ferrari kernel: [ 87.025308] RPC: Registered udp transport module.
Jul 14 01:49:05 Ferrari kernel: [ 87.025311] RPC: Registered tcp transport module.
Jul 14 01:49:50 Ferrari kernel: [ 131.552045] Clocksource tsc unstable (delta = -254345918 ns)
Jul 14 01:50:05 Ferrari kernel: [ 147.099483] rpcbind: server tsdsvr27 not responding, timed out
Jul 14 01:51:05 Ferrari kernel: [ 207.078987] rpcbind: server tsdsvr27 not responding, timed out
Jul 14 01:52:05 Ferrari kernel: [ 267.079721] rpcbind: server tsdsvr27 not responding, timed out
Jul 14 01:53:05 Ferrari kernel: [ 327.078024] rpcbind: server tsdsvrs02 not responding, timed out
The connection to the HP Hardy server is still up and working.
Other posts suggest a broken NFS package in Jaunty, I would prefer to keep this installation "clean" so that it can be duplicated to the other machines.

Question information

English Edit question
Ubuntu Edit question
No assignee Edit question
Solved by:
Martin Frauenstein
Last query:
Last reply:
Shang Wu (shangwu) said : #1

1. What version of NFS are you running on the IBM and on the HP?
2. Is there any difference in both /etc/exports ?
3. What's the error you get from the Jaunty client?
4. I assume that the other hardy client connects fine to both server, correct?
5. Please ensure that nfs-common is installed on Jaunty client.

Shang Wu (shangwu) said : #2

6. Please ensure that the Jaunty client can ping the IBM server too...

1. IBM server is running OS/400 v5R3. HP server is running Ubuntu Hardy (8.04).
2. Yes both /etc/exports are different as both servers are exporting folders from their own file systems.
3. The error is "mount.nfs: mount to NFS server 'tsdsvr27:/TSDData' failed: timed out, giving up" - tsdsvr27 is the IBM server. In the kern.log has an entry "rpcbind: server tsdsvr27 not responding, timed out"
4. Yes all other Hardy clients connect to both IBM and HP servers.
5. nfs-common 1:1.1.4-1ubuntu1, libnfsidmap2 0.21-2 and unfs3 0.9.21+dfsg-1 are the nfs related packages installed on the Jaunty client - Synaptic search of "nfs" on name produced the list.
6. The Jaunty client can ping both IBM server
64 bytes from tsdsvr27 ( icmp_seq=1 ttl=64 time=11.7 ms
and the HP server
64 bytes from tsdsvrs02 ( icmp_seq=2 ttl=64 time=0.238 ms

Entries in /etc/fstab are as follows:

# tsdsvr27 share
tsdsvr27:/TSDData /mnt/lan/shrtsdd nfs defaults 0 0
tsdsvr27:/GPT /mnt/lan/shrgpt nfs defaults 0 0
tsdsvr27:/TSDIntl /mnt/lan/shrtsd nfs defaults 0 0
## End of tsdsvr27 share

# tsdsvrs02 share
tsdsvrs02:/home/S02GPT /mnt/lan/shrs02 nfs rsize=8192,wsize=8192,timeo=14,intr
## End of tsdsvrs02 share

These entries are the same for the connecting hardy clients.

Shang Wu (shangwu) said : #5

1. Can you enable the debug on the IBM server and see if we can gather more information?
2. Can you perhaps try to run wireshark or other sniffer tool to see what might be the issue?

IBM server shows no errors.
Load wireshark and ran a test. Results saved and printed to text file. Results beyond my skill level. Posting text file would be too large. Can I email them to you?

Shang Wu (shangwu) said : #7

Sure. I will try to see what I can find in there.

Files attached

Shang Wu wrote:
> Your question #76911 on Ubuntu changed:
> Status: Open => Needs information
> Shang Wu requested for more information:
> Sure. I will try to see what I can find in there.

Regards & best wishes

Martin Frauenstein
Office : +64 (4) 920 6377
Fax : +64 (4) 920 6378
Mobile : +64 (21) 366 377
Skype : mfrauenstein
(JMJBWU) John 16:23-28 (14G27J36S5L)
PS: Your Silence accepts the contents of this message.

Shang Wu (shangwu) said : #9

I didn't receive the file in my email. Please try again to shang [AT] ubuntu [DOT] com

Ran a scan in wireshark. Noticed an entry from a strange local IP address. Found it on the IBM server. The physical adapter had been multi-homed with the "strange" IP address, which was no longer being used. Deleted the IP address and all is now working. BIG THANK YOU to Shang for suggesting Wireshark to scan for a possible problem and for the patience in dealing with me. THANK YOU.