[OpenIndiana-discuss] strange reboot behavior: deadman timeout

[Milan Jurik]

Hi,

On 21.02.2012 08:47, Lou Picciano wrote:
> All,
>
> Really stumped on this one: even after recent updates, we are unable
> to do a smooth software reboot; see a deadman timeout and CPU panic,
> followed by a core dump. Immediately thereafter, the machine _does_
> reboot itself. Hmmm....
>
> The machine has recently been upgraded to prestable-a1; prestable_a2
> is next. Prestable seems to have addressed the interrupt storm issue
> on this SandyBridge machine; vmstat 5 interrupt counts have dropped 
> to
> < 1k from 7+ _million_. So far, so good. But now we can't get past 
> the
> reboot issue.
>
> There may still be a timing issue after all - we are able to take the
> machine down occasionally with parallel gcc builds. Anyone have any
> suggestions/inputs on this?
>

few more questions:

1) "But now we can't get past the reboot issue."

- does it mean in previous releases you were able to reboot without 
problems?

2) reboot - does it mean poweroff is working?

3) which reboot does not work? fast or full?

4) "we are able to take the machine down occasionally with parallel gcc 
builds"

- does it mean system is crashing occasionally under load?

5) at which stage of system shutdown does deadman hit? Does it create 
crashdump?

> Regards,
>
> Lou Picciano

Best regards,

Milan