X1s do not have nvram the same way most other machines do.  They have
a rtc with a rather interesting programming interface.  Suffice it to
say that I have been unable to get any reasonable values in and out
of the rtc on that machine.

Anyway, even if you did manage to fix the time issues, the pciide 
driver seems to have problems with savecore.  It gets into a state
where the driver constantly times out and becomes practically
unusable even after savecore is terminated.  (One of these days
I need to get someone to decypher the output of the pciide driver
to determine the root cause.)

Eduardo

| I believe that I have worked this out, discovering that savecore does
| not like the date of the core file vs. the current date in nvram which I
| cannot seem to successfully set and retain on this X1. I've added a '-f'
| flag to $savecore_flags and will try that when I am not remote to the
| machine.
|
| > -----Original Message-----
| > From: Manuel Bouyer [mailto:bouyer@antioche.lip6.fr] 
| > Sent: Monday, May 13, 2002 10:15 AM
| > To: Randy Terbush
| > Cc: current-users@netbsd.org
| > Subject: Re: Problems getting a core dump.
| > 
| > 
| > On Thu, May 09, 2002 at 01:23:59PM -0600, Randy Terbush wrote:
| > > 
| > > I've been working on a very repeatable kernel panic that I 
| > cannot seem 
| > > to actually get a core deposited to /var/crash. I've forced 
| > the dump 
| > > in ddb before issuing a reboot. Any help would be appreciated.
| > 
| > What does savecore tell at boot ?
| > You should find it in /var/log/messages