Subject: Re: nfs servers and 5 minute VOP_READ's
To: Jason Thorpe <firstname.lastname@example.org>
From: Bill Studenmund <email@example.com>
Date: 03/15/1999 10:48:13
On Sat, 13 Mar 1999, Jason Thorpe wrote:
> On Fri, 12 Mar 1999 21:44:15 -0800 (PST)
> Bill Studenmund <firstname.lastname@example.org> wrote:
> > How badly would this break an nfs server? The first nfsd to read the file
> > would sleep for five minutes. What would happen then? Would the client
> > retry, and possibly send another process to sleep (possibly getting all of
> > them)?
> nfsd(8) itself doesn't actually do any reads/writes. What it does is
> set up the socket/authentication, and then run entirely inside the kernel.
> All reads/writes are done in the kernel.
Right, but they are done in the context of one of these servers, aren't
they? Thus if VOP_READ sleeps, one of these in-kernel server processes
will sleep, tying up one of the servers. ??