[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: snmpconf Comments on the BCP
There's no reason to believe, from the standards or from specific MIBs, that
any particular ordering of varbinds could be preferred (other than evidence
to back Mike's claim there are agents that do care, of course!). But such
mis-implementations should not be endorsed with any language that suggests
this is a "best" practice from an IETF point-of-view. If the best (or only!)
way to manage specific implementations/products uses a certain varbind
ordering then that is not a matter for IETF.
I'd suggest we have to delete the second paragraph quoted by Mike and change
the "should" in the first paragraph to a "must" (I assume lowercase is OK
here - this is not a normative document).
If people want to go further than this, replace the second paragraph with:
"In practice, some non-conformant agents may not correctly accept SETs with
certain orderings of varbinds: this document cannot recommend any best
practices for management of such agents."
From: email@example.com [mailto:firstname.lastname@example.org]On Behalf
Of Michael MacFaden
Sent: Monday, October 08, 2001 5:55 PM
Subject: Re: snmpconf Comments on the BCP
On Sat, Oct 06, 2001 at 01:19:52PM -0700, Andy Bierman wrote:
>At 03:18 PM 9/28/2001 -0700, David T. Perkins wrote:
> Section 4.5 is not really true. The order of the varbinds
>in the PDU is not important to agents.
This BCP is pretty short so I'll enclose it here:
4.5. Order of varbinds in a SET PDU
When a given SET PDU contains multiple varbinds, agents conforming to
the SNMP should be able to process the varbinds in any given order.
In practice, it is often preferable that management applications send
the varbinds in the order they are defined in the MIB module.
My experience writing mgmt apps w/sets against various
SNMP agents has been just the opposite.
If requested, I will post evidence to show various
products that don't support out of order varbinds.