NFS timeouts reverted to 1 second

It looks like recent changes have accidentally reverted NFS timeouts back to 1 second. The timeo value in the NFS mount setting is deciseconds (10 = 1 second).

type nfs4 (rw,noatime,vers=4.0,rsize=1048576,...,timeo=10,retrans=2,...)

This causes performance problems, especially with NFS over TCP:

http://www.eall.com.br/blog/?p=2338

We're back to seeing a very large number of NFS timeouts and LFS files have been written underneath mounts again.

I know we had intended to have this set to 5 seconds, but does that provide any real advantage over 60 seconds?