[ofa-general] Re: [PATCH 3/3] osm: QoS - parsing port names

Hal Rosenstock hrosenstock at xsigo.com
Thu Oct 25 06:32:06 PDT 2007


On Thu, 2007-10-25 at 14:56 +0200, Yevgeny Kliteynik wrote:
> Sasha Khapyorsky wrote:
> > On 06:49 Wed 24 Oct     , Hal Rosenstock wrote:
> >> On Mon, 2007-10-15 at 12:39 +0200, Sasha Khapyorsky wrote:
> >>> On 10:48 Mon 15 Oct     , Yevgeny Kliteynik wrote:
> >>>>>>  Switches have the NodeDescription filled by FW, and it's usually the
> >>>>>>  same string for all the switches.
> >>>>> It must not be same. Also I suppose that node description can be changed
> >>>>> at least for some managed switches even today.
> >>>>  Come on, man...
> >>>>  How many cluster administrators that you know will actually go and set
> >>>>  NodeDescription on switches???
> >>> I know at least one asked for this.
> >> Perhaps switch_map can be used in conjunction with this like in the
> >> diags ?
> > 
> > Hmm, right, switch_map is another example of switch naming, which is
> > useful with diags.
> > Perhaps even more generic - guid to name map? And this will work instead
> > of (or in addition to) node description when specified?
> 
> Can you elaborate on this?
> What exactly is switch_map?

See infiniband-diags (ibnetdiscover man page for one).

> And why would be need an additional guid-to-anything map if we
> already have node map indexed by guids (or am I missing something)?

GUIDs are not the most admin friendly identifiers.

-- Hal

> -- Yevgeny
> 
> > Sasha
> > 
> 



More information about the general mailing list