[CentOS] dell poweredge 860

Tue Feb 24 12:05:53 UTC 2009
William L. Maltby <CentOS4Bill at triad.rr.com>

On Mon, 2009-02-23 at 17:56 -0800, nate wrote:
> Jerry Geis wrote:
> > Hi all,
> >
> > Does anyone know of any issues with a dell poweredge 860?
> >
> > I have one and the NMI number is getting high.
> > I have a digium card in the box and the NMI grows with or without the
> > card in the box.
> 
> Is it causing a problem? I just checked a half dozen systems, all
> of them have pretty high NMI counters in /proc/interrupts(e.g.
> 254947480), and all of them are incrementing at least every
> second. No noticeable problems. Some of the systems are 3 years
> old, one of them is so new it's not even available for sale until
> next month.
> 
> nate
> <snip sig stuff>

I also believe that a high NMI count, in and of itself, is not an issue,
nor necessarily and indication of one.

IIRC, Non-Maskable Interrupts are ones that must be dispatched
immediately, not queued for handling. Hardware events that cause these
interrupts, IIRC, include the hardware clock, HDs, serial devices, etc.

Thinking just of the hardware clock alone you can see how NMI counts
would always be increasing.

I'd suggest doing what Nate did: look at some other systems to see what
happens to their counts over a brief time and see if yours looks
reasonable.

As also suggested, a better approach is the diagnostic software.
After-market vendors are notorious for crappy 1st level support. It's
not surprising the they suspect NMIs might be the problem. You might end
up spending some real time before the failure is properly identified and
rectified.

HTH
-- 
Bill