Re: [bug report] NFS and uninterruptable wait states

["Peter Wächtler" <pwaechtler@loewe-komp.de>]

Peter Wächtler wrote:
> 
> Phillip Susi wrote:
> >
> [ "mount -tnfs" with hard has default ]
> 
> > Anyhow, about an hour later ( the mount process still stuck ) I figured out
> > that the other machine was not running rpc.nfsd, though it was running
> > rpc.mountd.  Once I started rpc.nfsd on the machine, the mount on my box
> > finally returned ( and was terminated by the SIGKILL that I sent it an hour
> > before ).
> >
> > Could someone confirm that this is a bug, and explain why anything should
> > ever need to wait in that state?
> >
> Well, if you use the option "soft",then your process is interruptible.
> From a user standpoint, I don't understand the requirement of 'D' state.
> 
> Where this gets really impractical: commands like df or du will
> hang forever (if the other end is out of your control).

Err, "intr" will give you interruptible state. "hard" will retry forever,
"soft" will timeout and give up .

sorry