Thanks for the help. Unfortunately, this has not proved to be the solution
to my problem. Suddenly, however, one of the machines (still on autoneg,
btw) began working, without any warning or me touching it. All others are
still timing out on amdump (though amcheck still works).

--Justin

On Thu, 13 Feb 2003, Amanda Admin wrote:

> Justin,
>
> This sounds like the symptoms others on the amanda list have attributed to
> half/full duplex network interface and/or switch problems.
>
> I seem to recall a posting just recently saying that (the built-in eth
> interface on ??) Solaris had a particular affinity towards incorrect duplex
> detection. Maybe a search of the archives on this topic wwill turn up some
> details.
>
> Doug
>
> > -----Original Message-----
> > From: owner-amanda-users AT amanda DOT org
> > [mailto:owner-amanda-users AT amanda DOT org]On Behalf Of justin m. clayton
> > Sent: Thursday, February 13, 2003 2:37 PM
> > To: Joshua Baker-LePain
> > Cc: amanda-users AT amanda DOT org
> > Subject: Re: hosts timing out on amdump but not amcheck
> >
> >
> > On Thu, 13 Feb 2003, Joshua Baker-LePain wrote:
> >
> > > On Thu, 13 Feb 2003 at 9:15am, justin m. clayton wrote
> > >
> > > > First of all, thanks to all who helped me track down my NAK
> > problems from
> > > > last week. Having fixed that, all backup hosts pass amcheck
> > with flying
> > > > colors. However, when it comes time for the amdump, my log
> > report claims
> > > > "Request to <host> timed out" when I return the following morning.
> > > > However, if I run amcheck again, no hosts report problems.
> > This has been
> > > > going on for a number of days now. I am getting "Read error at byte
> > > > 0...:Bad file number" on some hosts (via
> > /tmp/amanda/sendsize.*), and some
> > > > are reporting "amandad: dgram_recv: timeout after 10 seconds
> > > > amandad: waiting for ack: timeout, retrying" (via
> > /tmp/amanda/amandad.*).
> > > > Strangely, though, the symptom is the same for all machines.
> > >
> > > What OS/distro?  Are there firewalls in the way?
> >
> > The clients are Solaris 8, the server is stable Debian linux, both using
> > 2.4.2p2. No firewalls in the way. This configuration has worked in the
> > past.
> >
> > Justin Clayton
> > VLSI Research System Administrator
> > University of Washington
> > Electrical Engineering Dept
> > justincl AT u.washington DOT edu
> > 206/543.2523  EE/CSE 307E
> >
>
>

Justin Clayton
VLSI Research System Administrator
University of Washington
Electrical Engineering Dept
justincl AT u.washington DOT edu
206/543.2523  EE/CSE 307E