[cgl_discussion] Serviceability reqs SMT 7.1 - 7.7

Mahesh Kallanagoudar mah206 at yahoo.com
Tue Dec 13 11:45:58 PST 2005


Corey,

Thanks for your response. 

CKRM provides monitoring for CPU, memory and network
bandwidth. So IPMI/CKRM should satisfy the
requirements, except for the weakness in IPMI event
reporting?

The other option is to use pcp (performance co-pilot).
But you have to write PMDAs for various boards you
would want to support. 
Also pcp doesnt interface with HPI. 

I would appreciate your suggestion on which approach
would be the best.

Also I found that OSDL has done some enhancements to
PCP. Is there any way I could download that?

Thanks
mahesh

--- Corey Minyard <cminyard at mvista.com> wrote:

> Mahesh Kallanagoudar wrote:
> 
> >Hi,
> >
> >It is specified in SMT7.1 - 7.7 that a frame work
> for
> >Diagnostics and Resource monitoring needs to be
> >implemented. Resource include temperature, voltage,
> >power etc.. 
> >
> >IPMI provides SDRs and Event logging/alerting for
> >these resources.
> >I am wondering why there is a separate requirement
> >apart from IPMI?
> >
> >Or does IPMI satisfy this requirement? What is the
> >need to implement  something like performance
> >co-pilot?
> >  
> >
> Resources also include things like free memory,
> ethernet bandwidth, and
> other things that IPMI has no idea about.  Plus the
> IPMI eventing
> mechanism is very weak for reporting anything but
> IPMI events.
> 
> -Corey
> 
> >I would appreciate a quick response.
> >
> >Thanks
> >mahesh
> >
> >__________________________________________________
> >Do You Yahoo!?
> >Tired of spam?  Yahoo! Mail has the best spam
> protection around 
> >http://mail.yahoo.com 
> >_______________________________________________
> >cgl_discussion mailing list
> >cgl_discussion at lists.osdl.org
>
>https://lists.osdl.org/mailman/listinfo/cgl_discussion
> >  
> >
> 
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 



More information about the cgl_discussion mailing list