Re: RAIDframe swap problems?

[Chris Ross <cross+netbsd%distal.com@localhost>]

On May 5, 2012, at 20:23 , matthew green wrote:
>>  Anything else I can provide for diagnostics?
> 
> two things to try:
> 
> - ktrace savecore.  what does it do?

  I did this.  It looks around a lot, and eventually opens ioctl's
(KIOCGSYMBOL) /dev/ksyms a few times, then reads data
from  /dev/rraid0b via pread syscall.  I also notices that most of
the ktrace is full of stat50, lstat50, fstat50 calls.  What are those?
They read like old syscalls, but this is a new system...

  Anyway, after it reads a bunch of data from /dev/rraid0b, it goes
into looking at timezone and posixrules things that I assume are
related to generating the error message.

> - manually zero the dump partition with dd, see what happens.

  I did this.  After doing it, the savecode then ran successfully and
said "savecore: no core dump".

  Worth chasing, or should we just assume it not worth chasing
down this "bug"?

          - Chris

ps, I have the ktrace outputs for the savecore and savecore -c runs
before I zero'd the partition, if anyone wants the full data.