NFS mount timeout suspicion

[Scott McDade]

Hello:

I am trying to troubleshoot a very sporadic and frustrating problem. It looks like a problem with my NFS mount but I can’t say for sure because I don’t know what logs or utilities to use to help me understand the issue. I believe my NFS mount point is timing out but I need to be able to prove it. Can anyone suggest any logs or utilities I can use to trouble shoot this problem? On the client side I have a Red Hat Linux Desktop 5.3 and on the sever is a NetApp filer. It may simply be how the NFS mount points are configured and I need to be able to accommodate high network activity without timeouts.

[Matti_Kurkela]

Are you soft-mounting the NFS filesystem?

If you do that and are suffering from timeout error, you might wish to switch to hard mounting. Soft mounting is generally not recommended for writeable filesystems, nor for filesystems that contain executables.

Change the NFS mount option "soft" to "hard,intr" on the client, and re-mount the NFS filesystem(s). The "intr" option allows you to interrupt an application that is stuck waiting for NFS I/O if you wish, with the regular kill command (or Ctrl-C).

http://www.brandonhutchinson.com/nfs_timeouts.html

If your filesystem is hard mounted and connection to the NFS server fails for any reason, the client will keep retrying and will output the classic "NFS server X not responding, still trying..." error message. These messages will be produced by the kernel, and the backlog of the most recent kernel messages is always available with the "dmesg" command.

Any important messages should also be recorded to log files in /var/log: check your /etc/syslog.conf or /etc/rsyslog.conf to see which log file is set to receive messages of the "kernel" category, and adjust the priority level of recorded messages if necessary.

MK