I reverted my sysctl settings and indeed it stalled again when using your cp-script. After reenabling the sysctl settings and copying about 40GB I still could not get it to stall. So it seems that the sysctl settings really fixed the problem for me. At least with Lucid kernel 2.6.32-24.42.
For comparison here are my NFS settings. All clients are connected via GB-LAN:
@Timo Harmonen
I reverted my sysctl settings and indeed it stalled again when using your cp-script. After reenabling the sysctl settings and copying about 40GB I still could not get it to stall. So it seems that the sysctl settings really fixed the problem for me. At least with Lucid kernel 2.6.32-24.42.
For comparison here are my NFS settings. All clients are connected via GB-LAN:
Server: rw,sync, no_subtree_ check,no_ root_squash rsize=16384, wsize=16384, acl
Client: hard,fg,
I hope you'll find a solution for your problem, too. Good Luck.