Re: [Freeipmi-users] ipmiconsole asus problem

[Albert Chu]

> Hello Al,
>
> On Sunday 23 March 2008, Albert Chu wrote:
>> Hey Bernd,
>>
>> > Hello Al,
>> >
>> > On Sunday 23 March 2008, Albert Chu wrote:
>> >> Hey Bernd,
>> >>
>> > I still have a problem, though. As with ipmitool sysrqs don't work,
>> > although
>> > sysrqs are generally enabled. Do you have an idea what might be the
>> > problem?
>>
>> I can't speak for the asus boards, but I know that some motherboards
>> cannot support serial-breaks via IPMI, therefore not allowing sysrqs to
>> be
>> generated.  I only know of one vendor that does support it, and I had to
>> complain until a firmware fix was finally done.  So it wouldn't surprise
>> me if serial-breaks via IPMI just didn't work on the asus board.  Do
>> sysrqs work when you connect via a normal serial cable?
>
> I would be suprised if sysrqs via serial cable wouldn't work, so far this
> was
> never an issue with any system I ever tested. But I will make absolutely
> sure
> on Tuesday when I'm on site.
> We have an ipmi stability problem anyway and Asus is presently designing
> new
> ipmi cards (AFAIK some internal memory of these cards is too small to
> cache
> all sensors data and then the cards simply crash after some time), so I
> will
> tell them about the break problem...
> (I really need sysrqs, since I'm actually responsible for Lustre and the
> day
> will come when I need kernel traces to debug a Lustre problem).
>
> Thanks again for your help and happy Easter,
> Bernd
>

Hey Bernd,

Thanks for working with Asus on it.  We use sysrqs here all the time for
kernel traces/crashes/etc.  The lack of its functionality by a lot of
vendors in IPMI has worried us.  Getting each vendor to be more IPMI
compliant and get IPMI better will be good for all of us.

Al

-- 
Albert Chu
address@hidden
925-422-5311
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory