Re: shutdown -p panics on SS10

On Tue, 13 Jun 2023 at 11:52, Riccardo Mottola
<[email protected]> wrote:
>
> Hi David,
>
> David Brownlee wrote:
> > How easy is it to trigger the netbsd9 issues?
>
> Not easily... the only way I found to reproduce is to keep the machine
> running and working for some time... usually one or two days compiling
> stuff and being connected through ssh.
>
> > As a datapoint, would it be worth trying to extract a netbsd-9 kernel
> > (as /netbsd9) & modules onto the netbsd-8 and try a boot into the
> > netbsd9 kernel (retaining everything else) to see if the issue
> > exhibits?
>
> so netbsd 9 kernel on 8 userland on the SS10 ? Could be worth a test,
> but as said I have no "safe and certain way" to reproduce the issue,
> which makes it hateful.

They are always the most annoying issues. I was looking for something
relatively easy to do, but more importantly, easy to revert :)

> My SS10 had issues when trying to boot kernel outside 1GB, I didn't make
> a separate parittion back then when I upgraded the HDD to the "mind
> bogging" 4GB :) That was back in NetBSD 2.0 times! It was also very
> fincky about accepting a second HDD. With various tests of "known good"
> HDDs I ended up breaking the cable. I found a replacedment, but then
> stopped "playing around" with this nice machine. I'll check how it is
> configured a the next boot.

Another thought might be to copy the good netbsd-8 machine across to
the problem machine to see if the issues still show (albeit at the
temporary loss of netbsd-9 there :)

Do you have them setup to netboot? Quite a while back I supported a
student(ish) cluster of mixed sparc boxes, some of which had SCSI
drives that failed but could be recovered with a reformat, and with a
spiderport console I had the process pretty much automated to
reformat, reinstall and reconfigure once I manually netbooted (from
eight timezones away at times :)

David