Subject: Re: IP when date is wrong
To: der Mouse <mouse@Rodents.Montreal.QC.CA>
From: Greg Ingram <email@example.com>
Date: 01/18/2001 09:08:20
On Thu, 18 Jan 2001, der Mouse wrote:
> > I thought my DEQNA ethernet adaptor died after a recent power outage.
> > The kernel properly detected and configured it but I couldn't ping
> > anybody. It turns out that the date was wrong (0) and when I set the
> > correct date IP traffic started flowing again.
> > I suspect this isn't a VAX-specific problem, but is it proper
> > behavior or is it a bug?
> In my experience, if the date is 0 then the clock chip isn't running at
> all, and in that case a whole bunch of timeout-based things seem to
> fall over. I've never tried doing networking when in that state, but
> it wouldn't surprise me if clock interrupts weren't happening and hence
> soft interrupts (soft network interrupts in particular) weren't
> happening - which would break much of networking.
> This "experience" is on a uVAX-II. (You mention a DEQNA, but that
> merely means a Qbus VAX, not necesarily the uV2.)
It is a uVAX-II. I'd think clock interrupts are happening because I only
ran 'date ...' and now it's keeping time, etc. Hmm....