> Just for starters, I don't understand the difference between nsCPUIndex and
> nsHRCPUDeviceIndex.
Well, I thought it was worth being able to link together the net-snmp
specific table, with the standard hrDevice table. So I wanted *some*
mention of hrDeviceIndex here.
The two options were:
a) make this table a (sparse) AUGMENTation of the hrDeviceTable
and hence indexed by hrDeviceIndex directly
b) make this table indexed separately, but include a column object
to map into the hrDeviceTable
I'd actually prefer the first, but wasn't sure how generally acceptable
this would be. Particularly since the HR-MIB wasn't (and still isn't)
part of the default configuration.
> On one of my V880's it identifies its processors (in sequence) 6,4,0,1,2,3
> and this sequence appears to have (albeit inscrutible) meaning. Could we
> therefore have a separate ID field? In addition, I'm able to report
> additional parameters for CPU's including description, speed and
> online/offline (status). Rather than have these scattered between this data
> structure and HOST-RESOURCES-MIB, woulnd't it be easier to read if we had
> them cobbled up into one?
My gut reaction is to avoid duplicating information, so I'd rather
use the HR-MIB objects for that. Which might be an argument for
having this table AUGMENT the hrDeviceTable.
> You had mentioned "An index of 0 should be reserved for aggregate statistics
> on a multi-CPU system." so I'm not sure what to do in that case with my
> extra fields.
If the fields aren't relevant for a particular row, then they should be
empty - i.e. return 'noSuchInstance'
IMO, of course.
I'd be interested in hearing other views....
Dave
-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
|
