Re: NetBSD 5.1 RC3 in production

[buhrow%lothlorien.nfbcal.org@localhost (Brian Buhrow)]

        Hello.  If you're using raid1, then you can just have the kernel dump
to one of the swap partitions explicitly.  I've gotten some good kernel
dumps that way under NetBSD-5.  Then, when you run savecore, I've found you
have to run with the -N flag explicitly to get it to pull the core file off
of the swap partition after the fact.  This means, the default
/etc/rc.d/savecore script won't work out of the box, so be prepared to log
in shortly after the thing reboots, or, at least before it starts using
swap, and run savecore manually.
-Brian
On Sep 20,  2:17am, Robert Elz wrote:
} Subject: Re: NetBSD 5.1 RC3 in production
}     Date:        Sun, 19 Sep 2010 14:12:41 -0400
}     From:        Thor Lancelot Simon <tls%panix.com@localhost>
}     Message-ID:  <20100919181241.GA16013%panix.com@localhost>
} 
}   | Why not hook up a serial console and just look at what the kernel logs?
} 
} At the minute I have nothing convient that I could connect the other end
} of the console cable to - munnari is kind of a loan system.   But I'll see
} if I can make that happen if the problem persists and nothing else provides
} any info.
} 
}   | Are you not getting kernel cores?
} 
} No, swap is on raidrframe (everything is on raidframe), and I don't think
} dumps work to raid do they?    I'll see if I can find some empty non-raid
} space and make that be dumpdev.
} 
} kre
} 
>-- End of excerpt from Robert Elz