[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: snmpconf BCP-07 vs configuration management



On Wed, Nov 28, 2001 at 04:28:06PM -0800, rpresuhn-lists@dorothy.bmc.com wrote:
>However, it still doesn't do justice to some of the
>central issues of configuration management.  (as defined in
>http://www.itil.co.uk/online_ordering/itil_glossary.htm ) For
>example, though the document spends quite a bit of time on the
>transaction concept, only one sentence in section 5 (2) seems
>to even touch on the issue of versioning of configurations,
>and almost nothing on the concept of identifying configurations.

I agree versioning configurations is a very important part 
of the configuration process. In my experience, that has
been done outside of scope of the Internet Standard
Managment Framework and is very well understood.

We BCP editors did discuss how difficult it is to get a configuration
off a device using SNMP, and that writing a generic configuration
extraction tool for an arbitrary device is impractical. 
Jon did mention to me a nice solution to this problem but 
it has not been defined yet as a standards track MIB module. 

>I think it would be worthwhile to provide a higher-level
>view by including more of the concepts from ITIL, at least
>by reference.  Otherwise, it has more of the flavor of a
>tutorial on how to implement SNMP, rather than a compendium of
>best current practice for the use of SNMP for configuration
>management.  Does the best current practice really have so
>far to go?

My goal was enumerate good and contrast with not so good
implementation mechanisms and processes that directly involve
components of the management framework. 
If this level of knowledge is not achieved and implemented,
then any higher level view of configuration won't be of much use 
to the practitioner. I'm not against adding some references to 
some higher model(s), it is just not the case everyone subscribes 
to the same models.  

For example from your glossary: "Expert User" See 'Super User' 
Give me a break.  'Super user' is just a particular account on a *nix
box right? (I'm not kidding). And I've never seen 
'Super Users' be required to do first line support
in an organisation that wants to retain them. 

>I propose adding the references to the ITIL definitions
>to section 1.2, and extending section 2 by adding another
>subsection dealing with identification and versioning of
>configurations using SNMP.  

I just haven't seen any systems in practice given the scope
described above. For CLI methods of configuration, this 
is the one pattern I've seen used most:
   http://www.shrubbery.net/rancid

>Section 5 should be expanded
>so that point (2) in the first paragraph is sytematically
>addressed, particularly since this can interact with the
>persistance / activation mechanisms discussed in 3.7+

Maybe the sample MIB module could be used to provide
such a secenario?

Regards,
Mike