Server reboots at random :-(

[Barry Friedman]

On Fri, Jun 20, 2008 at 3:50 AM, Ulf Zimmermann <ulf at alameda.net> wrote:
> On Thu, Jun 19, 2008 at 07:29:41AM +1000, Edwin Groothuis wrote:
>> On Wed, Jun 18, 2008 at 12:29:00PM -0400, BarryFriedman at www1.emax.ca wrote:
>> > Any ideas about how to troubleshoot this would be greatly appreciated.
>>
>> Start with enabling crash dumps to see if it is a kernel issue or
>> a hardware issue:
>
> You might also want to setup virtual serial and enable console on serial.
> Then you can use conserver to log anything and check after a reboot.
>

I am trying to get virtual serial boot console working with ilo standard
on the DL380G4.  First off, I can't see any way to get the ILO to
reconfigure and use COM1 as the virtual port. Anders Nordby had posted
a similar problem in Oct '07 in which he apparently was configuring the
RBSU from the command line. Was this using ilo2 or advanced ilo? In any
case, the alternative is to redirect the console to COM2 and reconfigure
the boot parameters. The BSD handbook says to use the line : 

  device sio1 at isa? port IO_COM2 flags 0x10 irq 3

to do this, but config reports a syntax error on this line. Is the
handbook in error? Do I need to set the flags for sio1 in the HINTS
file? What is the correct way to proceed?

The reboot problem persists. The machine has gone as long as 27hrs
before rebooting, average 18hrs. This is a system under no load. I have
tried turning off the network interface to see if there was any external
source of the problem to no avail.  What kinds of hardware failure could
cause this kind of behavior?


Regards,
-- 
Barry Friedman                         
Emax Computer Systems Inc., 480 Tweedsmuir Ave., Ottawa, Ont. Canada K1Z 5N9
bfriedman at emax.ca    Phone: (613) 725-3198  Fax: 725-0298