Re: [802.1] MIB definitions for Link Aggregation

[Michael Kirkham <mikek@muonics.com>]

Something else I noticed that is possibly wrong with this MIB, from
running it through our compiler:

dot3adTablesLastChangedGroup OBJECT-GROUP
    OBJECTS {
        dot3adTablesLastChanged
    }
    STATUS   current
    DESCRIPTION
        "A collection of objects providing information about the
        time of changes to the configuration of aggregations and
        their ports."
::= { dot3adAggGroup 6 }

This was probably intended to be { dot3adAggGroups 6 } as with the other
groups defined within the module.  Probably a typo, but also a lot of
compilers will complain about (and some will be unable to work with) the
fact that a group is defined subordinate to another group.

This module appears to have been last updated 3 years ago, and not being
familiar with it I don't know if this is a "known" thing and people have
implemented it as dot3adAggGroups.6 or if (likely) it's stuck the way it
is due to revision rules (which would require a different descriptor for
the group if fixed, as well as new descriptors and OIDs for any compliance
statements that reference the group).  I've copied the listed contact for
the MIB so they are aware of the 802dot3 issue as well as this item, if
they're not already, and can make any changes they deem appropriate.

I suspect the issue below was probably brought up because someone's MIB
compiler rightly complained about '802dot3' but didn't give a sufficiently
clear error message.  The issue below is one of legality, the issue above
is one only of compatibility.

On Wed, 14 May 2003, Michael Kirkham wrote:

> Date: Wed, 14 May 2003 03:30:57 -0700 (PDT)
> From: Michael Kirkham <mikek@muonics.com>
> To: Les Bell <Les_Bell@eur.3com.com>
> Cc: mibs@ops.ietf.org
> Subject: Re: [802.1] MIB definitions for Link Aggregation
>
> On Wed, 14 May 2003, Les Bell wrote:
>
> > Hello list,
> >
> > I have a doubt about the MIB recommonded in IEEE 802.3, 2000
> > appendix
> > 30C SNMP MIB DEFINITIONS for LINK AGGREGATION.
> >
> > I copy-pasted the relevant information here for your reference.
>
> A reference to the whole MIB would probably be helpful for those of us not
> familar with the module in question.  I'm assuming this is the correct
> link:
>
> http://grouper.ieee.org/groups/802/3/publication/ad/IEEE8023-LAG-MIB.txt
>
> >        ::={iso member-body(2) us(840) 802dot3(10006)
> > snmpmibs(300)
> >            linkagg(43)}
>
> In the version referenced by the above URL, linkMIB is defined with an OID
> that doesn't include a descriptor for the last subidentifier, but is
> otherwise equivalent:
>
>     ::= { iso(1) member-body(2) us(840) 802dot3(10006)
>           snmpmibs(300) 43 }
>
> These are the same OID.  The "linkagg" in "linkagg(43)" doesn't actually
> mean anything.  It doesn't define or register anything.  It's only for the
> reader's benefit.  Without seeing some other MIB module registering
> 1.2.840.10006.300.43 to some other definition, I would assume that the OID
> is correct as it appears.
>
> A more interesting thing of note, however, is that "802dot3" is illegal.
> Even though those descriptors don't define anything, they still must be
> valid identifiers, starting with a lowercase letter, if this
> "NameAndNumber" form is used.
>
> --
> Michael Kirkham
> www.muonics.com
>
>

--
Michael Kirkham
www.muonics.com