Problems mounting nfs

[Carles Viaplana]

Hello,

I have problems mounting an nfs filesystem from a remote HP-UX to a Linux server:

budapest:/ # mount osiris:/INTERFACES /INTERFACES/
nfs server reported tcp not available, falling back to udp
Using NFS over UDP can cause data corruption.
Please refer to the WARNINGS section of the nfs(5) manual page.

I already mounted other nfs from same remote server to local server, but I don't know why I can't with this directory.

Any idea?

Thanks in advance for your help.
Regards,

Carles

[Carles Viaplana]

Hello again,

I checked processes and I've found these:

budapest:~ # ps -ef | grep INTERFACES
root 18465 1 0 11:58 ? 00:00:00 mount osiris:/INTERFACES /INTERFACES/
root 18665 1 0 12:00 ? 00:00:00 mount -t nfs osiris:/INTERFACES /INTERFACES
root 18929 1 0 12:05 ? 00:00:00 mount osiris:/INTERFACES /INTERFACES/
root 19847 19312 0 12:25 pts/2 00:00:00 grep INTERFACES

I tried to kill them but I can't. Any idea why?

Thanks in advance for your help.
Regards,

Carles

[Carles Viaplana]

Hello again,

I tried to mount nfs from other HP-UX systems and I'm able.

Does it means problem is in nfs configuration from remote system?

Thanks in advance for your help.
Regards,

Carles

[Ivan Ferreira · Accepted solution]

From linux can you post the output of:

rpcinfo -p nfs_server
showmount -e nfs_server

Do you have some firewall between the server and the client?

[Carles Viaplana]

Hello,

Here you're the outputs:

budapest:~ # rpcinfo -p osiris
program vers proto port
100000 4 tcp 111 portmapper
100000 3 tcp 111 portmapper
100000 2 tcp 111 portmapper
100000 4 udp 111 portmapper
100000 3 udp 111 portmapper
100000 2 udp 111 portmapper
100024 1 tcp 53314 status
100024 1 udp 60692 status
100021 1 tcp 53315 nlockmgr
100021 1 udp 60697 nlockmgr
100021 3 tcp 53316 nlockmgr
100021 3 udp 60698 nlockmgr
100021 4 tcp 53317 nlockmgr
100021 4 udp 60699 nlockmgr
100020 1 udp 4045 llockmgr
100020 1 tcp 4045 llockmgr
100021 2 tcp 53318 nlockmgr
100005 1 udp 60704 mountd
100005 3 udp 60704 mountd
100005 1 tcp 53319 mountd
100005 3 tcp 53319 mountd
budapest:~ # showmount -e osiris
Export list for osiris:
/INTERFACES (everyone)
/usr/sap/put (everyone)
budapest:~ #

Thanks for your help.
Regards,

Carles

[Ivan Ferreira]

I will try the following:

- First, clear all mount commands pending, reboot the client if needed.

- Second, ensure that name resolution is consistent, this could cause the problem. Check DNS configuration or configure the /etc/hosts file to include the hostname and ipaddress of the server and the client.

- Run the mount command, check with netstat the status of the connection at server and the client. Post the results

- If you cannot mount yet, we will use tcpdump to try to identify the problem.

[Carles Viaplana]

Hello Ivan,

I rebooted system but I'm still getting same error.

What do you propose?

Thanks for your help.
Regards,

Carles

[Carles Viaplana]

Just one question:

A colleague told me some kernel parameters from remote host were modified.

Could these changes affect nfs connections?

Regards,

Carles

[George Liu_4]

DO you any chance to restart nfs service on the HP server?
Also, any security restrictions, such as in firewall, tcp-wrapper?
What's the output of cat /proc/mounts on the linux client?

[George Liu_4]

For the quesition after kernel parameter change. May or may not. All depends on what parameters have been changed.

[Ivan Ferreira]

1- Check the messages in the HP-UX syslog.
2- Start tcpdump:

tcpdump -i -nlf host nfsclient and nfsserver

3- Post the output will see what is happening.

[Carles Viaplana]

Finally I found solution!

Problem was someone stopped nfs server on HP-UX system just before I tried to mount nfs from linux system.

Thanks to all for your help.
Regards,

Carles