suddenly cannot mount nfs share from Windows 7

Question

I recently reinstalled my file server (moved from fedora to Ubuntu server).

Now I cannot mount my nfs share from Windows 7, mounting from Mac OSX works fine.

In Windows I either keep getting "the semaphore timeout period has expired" or "an unexpected error has occured".

Does Ubuntu need some special magic to allow Windows 7 to mount an nfs share?

This is my exports file

/home/Bing/     192.168.1.*(rw,async,insecure,no_subtree_check)
/home/Bing/mnt/EXTRN2 192.168.1.*(rw,async,insecure,no_subtree_check)
/home/Bing/mnt/EXTRN3 192.168.1.*(rw,async,insecure,no_subtree_check)

Answer 1

Ubuntu might defaults to a newer NFS version than your old Fedora did. If Ubuntu already supports NFSv4 by default, then I don't know how Windows 7 handles that one. You might like to try to append nfsvers=3,mountvers=3 flags to your exports lines and see if that fixes your problem.

Answer 2

There are actually several programs that implement NFS. Is your rpc.statd and lockd running when you type in ps aux? If rpc.statd doesn't run, the connections will not work properly. Look in the /etc/init.d directory for another script that has "nfs" in it; most likely you'll need to start that script too.

Answer 3

Apply the following pseudo-patch to /etc/default/nfs-kernel-server:

-RPCMOUNTDOPTS=--manage-gids
+RPCMOUNTDOPTS=

I encountered the "the semaphore timeout period has expired" problem on Windows 7 while trying to access NFS on Debian Squeeze, so this is probably applicable (due to Debian being upstream for Ubuntu and the proximity of the Squeeze and Lucid release dates), but possibly not.

I could not replicate the problem accessing NFS on (default installs of) RHEL 6 or SLES 11, so upon comparing the running processes related to NFS (ie rpc.*) I noticed the addition of "--manage-gids" to rpc.mountd on Debian Squeeze, retried NFS running rpc.mountd without "--manage-gids" with success, and tracked that argument to /etc/default/nfs-kernel-server.