On Sun, Mar 17, 2002 at 06:46:07PM +0100, Ignatios Souvatzis wrote:
> 
> yes. As you're not waiting for the server to be reachable again (e.g, to
> reboot), but timeout and get an error.

OTOH if you have mounted a filesystem that doesn't persist across reboots
(it might be exported by hand) , or one where the 'file id' or 'filesystem
id' isn't guarateed to be the same after a reboot (most union fs, or
inode free fs have problems here) the you can end up with infinite
retry requests after a server reload.

The only way to get rid of the hard mount is to reload the client
- since the filesystem isn't (ever) going to come back.

I dunno what happens in netbsd, but on a different (unix) OS there
were no delays in the retry loop (at the client or server).  We
used to get over 1000 retries a second (on 10M ethernet)  when the
server reloaded.

	David

-- 
David Laight: david@l8s.co.uk