Technical article 112823

My server panics with an NMI (Non Maskable Interrupt) Error Message.

OpenServer 5.0.0OpenServer 5.0.2OpenServer 5.0.4OpenServer 5.0.5OpenServer 5.0.6OpenServer 5.0.7UnixWare 7.0.0UnixWare 7.0.1UnixWare 7.1.0UnixWare 7.1.1UnixWare 7.1.3UnixWare 7.1.4Open UNIX 8

Written · Last revised

Problem

My server panics with an NMI Interrupt Error Message, for example:

PANIC: An unknown NMI interrupt occurred.
eax: C082A000 cbx: C082C648 ecx: C0450058 edx: 0
efl: 246  esi: C082C678 edi: 246
esp: FFFDABC0 ebp: FFFDABC4  regset: 0
mps_idle_self+22: pushfl

Cause

You have a hardware problem on your server.  In the example
messages above, the Multi-Processor kernel has received an NMI
interrupt from a piece of hardware.  This kernel has not been
able to determine the exact location of the fault.

The "mps_idle_self" function is generated from the Multi-Processor kernel (mps) module:

cd /etc/conf/pack.d
for i in *
do
        echo $i
        nm $i/Driver* | grep "mps_idle_self"
done

The fault is hardware related but it is possible that a Single Processor kernel will not detect the fault, in this example.

Solution

The solution involves a step-by-step approach to try to troubleshoot the location of the problem:

1. The NMI Interrupt could have come from the motherboard, the memory, from a data bus, from a PCI or ISA card, a defective processor (or a mix of processor revision).

The first step would be to run some hardware diagnostics that should be available from the server's manufacturer.

Sometimes, simply re-seating the processor, memory and cards is sufficient to overcome this problem.

It may, though, be necessary to replace the server's memory, CPUs, boards and motherboard one by one until the problem is rectified.

2. There may also be some hardware diagnostics from the network card manufacturer.

3. Check the firmware revisions of the server and any host bus adapter (SCSI) cards from the manufacturer.

4. Check the hardware drivers available for your host bus adapter cards or network cards from https://ftp.scoexpert.com/<os>drivers/ or if the driver isn't there, check with the manufacturer.

See http://www.sco.com/chwp for details of supported hardware.

5. Check the patch revisions of your operating system by visiting
https://ftp.scoexpert.com/<os>; & view the README file there for your
operating system.  Make sure the latest Intel Pentium Microcode
patch is installed, if your platform is Intel.

6. Check with the manufacturer for any new Platform Support Modules (PSM) for Multi-Processor support for the UnixWare 7 operating system.

Further analysis of the dump that UnixWare 7 creates will also
help identify the problem hardware.  The dump should identify
a consistent memory address, unless the hardware is changed.
NOTE:     If there is an "eip" entry in the panic dump, if the reference to it
          is consistent when the server panics then it is possibly a driver
          issue, otherwise if the "eip" entry changes with every panic then it
          is possibly a hardware issue.

See also

ta112001, "How can I force a dump on UnixWare 7?"

The Intel Developer's Web Page:
http://developer.intel.com/

From the Xinuos technical article database, record 112823. Reproduced with permission. The original is at www.sco.com/ta/112823, which is reachable but blocked from search engines.

Next step

Need this fixed rather than explained?

Tell us the release, what the system runs and what it is doing. We will reply by email with what we would do next.

  • You deal with the people who do the work
  • No obligation, and no upgrade required before we will help
  • We tell you plainly if something is not worth doing

Send an enquiry

We reply by email. If a server is down, say so in the first line. You can also write to help@scoexpert.com.