[SunHELP] Dump device on mirrored swap

[Nick Steel. (lists)]

Hi Folks -

A couple of days ago, I had a system running 2.7 panic. It left a core in
/dev/dsk/c0t0d0s1 which is the dump device listed in /etc/dumpadmp.conf.
However, when the system rebooted, savcore reported that the dump had
already been processed. I tried a savecore -d -f /dev/dsk/c0t0d0s1 and it
reported "bad magic number 0" which indicated that there wasn't a core
present.

I'm not quite sure what happend here, my console buffer log shows that the
system did finish writing the core to the dump device, but savecore won't
pull the core out. I did a bit more digging, and this system is using Disk
Suite to mirror the slices on the boot disk, including primary swap.
c0t0d0s1 is one of the slices used to create the meta device which is used
for primary swap.

So, my question is, if the system write a core to one of the slices that's
used to create a mirror under Disk Suite, would Disk Suite try to resync the
two slices in the mirror and overwrite the core? If that's not the case,
does anyone have any ideas as to what happend with the core that got written
out?

Thanks!

- Nick