I find the NFS timeout in the netmount init script to be a bit long (I just rebooted with the NFS server down, and I estimate that it took about 10 minutes for it to give up). The init script should use a lower timeout - 10 to 30 seconds would be much better. Reproducible: Always Steps to Reproduce: 1. 2. 3.
This time-out problem now seems to be resolved on my box. before I had 'sleep 2' to fix, now, with the newer nfs init.d script just installed today, it appears to bypass 'sleep' entirely now. ... as it should i would think. thanks for acknowledging/adjusting ;-) oh, and reading over you bug post, i believe the old init.d script was hanging my box on boot for ~30 (which corresponds to teh sleep statement within the init.d script).. and was unnecessary for a common nfs server (i'm guessing). now, w/ the adjustments, the script recognizes that it doesn't have to sleep at all.
I'm sure the problem was in nfsmount, though, and that only starts statd with start-stop-daemon (how long can that take?) and executes mount -a -t nfs... which must be where the problem lies.
It was a bug in the dep caching code that did not pull in the portmap and nfs rc-scripts.