Ok, that answers the first part of my question in regards to loading the
MIBS. Now, how do I make them useful, i.e. I'm receiving undefined traps
from my cisco agents. How/where do I define them, and make them
recognizable.
Thanks,
Jeff
>From: James Shanks <James_Shanks@TIVOLI.COM>
>Reply-To: Discussion of IBM NetView and POLYCENTER Manager on NetView
> <NV-L@UCSBVM.ucsb.edu>
>To: NV-L@UCSBVM.ucsb.edu
>Subject: Re: How to load and configure cisco mibs and traps?
>Date: Thu, 2 Mar 2000 17:15:42 -0500
>
>Ken,
>
>Perhaps you, or someone else (don't mean to put you on the spot) can
>clarify
>this for me. I have heard these sorts of remarks before, but I don't
>understand
>what exactly is meant by them. The mosy compiler, as I understand it, is
>freeware off the web, correct? Well, I don't have one and I have no idea
>what
>levels are available. No one I know has a copy of it.
>
>NetView does not ship the mosy compiler nor any other compiler but its own,
>so
>far as I can see. The code we have may have been based on the mosy
>compiler
>once upon a time, but that code has been modified many times since then and
>is
>internal to our product. It links to our libraries and I can see the
>tracks of
>many former IBMers (and even older HP employees) in it. We receive no
>updates
>for this code from any other source, and if it is broken, then we fix it
>ourselves without regard to code provided from any other source. I have no
>way
>of telling what level of the mosy compiler we may have started with, if
>indeed
>that was where we started.
>
> Most of the problems I see with vendor MIBs today are because they test
>them
>with the SMICng compiler (which is also free off the web) and it is a
>multi-pass
>compiler -- it builds a copy of the user's MIB and will even correct
>statements
>in it, adding spaces and punctuation if required, provided the errors are
>minor
>, before it starts the compile. There are a whole host of flags which can
>be
>set to have it tell you about what changes it is making and it will give
>you
>warnings about what may not fly with some other compilers, which are
>single-pass compilers, such as NetView's. But I don't think many vendors
>are
>being careful about this.
>
>The state of the MIB compilers in NetView at this time (5.1.2) as I see it
>is
>this:
>We have two, a V1-ony compiler (xnmloadmib) and a V1/V2 compiler
>(xnmloadmib2).
>
>The V1-only compilier is called in dialogs within the ruleset editor, and
>within
>snmpCollect's configuration dialogs, and its binary library
>(/usr/OV/conf/snmpmib.bin) is loaded by trapd, snmpCollect, and netmon.
>Once
>MIBs have been loaded with the V1-only compiler they can be browsed with
>its
>matching browser, xnmbrowser. A quick check of open defects shows that
>that
>there are no open defects with the V1-only compiler, although one was
>closed
>recently as a permanent restriction regarding the fact that white space is
>required between major MIB constructs (for example "XYZ ::= OCTET STRING"
>will
>compile fine whereas "XYZ::=OCTET STRING" will not because the compiler
>depends
>upon spaces to parse). All SNMP V1 MIBs should compile fine with this
>compiler, provided they have adequate white space.
>
>The V1/V2 compiler (xnmloader2) will parse V1 MIBS and V2 MIBs, but it is
>not
>linked to any other area of the product at this time except for its
>matching
>browser (xnmbrowser2), so it is of very limited use. As a result, many
>issues
>or defects with it remain unresolved at this time, though just about every
>maintenance release of NetView since 5.0 has made improvements to this
>compiler.
>What issues remain? Well, development on it stopped at the draft standard
>for
>SNMP V2, so items added to the standard after that are not well-supported.
>Currently, the AGENT-CAPABILITIES construct and BITS construct remain
>unsupported, and there are some issues with how it handles labels and
>matching
>levels, but there are defects open on these already and they are being
>pursued.
>
>Development is well-aware of the current shortcomings with the V1/V2
>compiler
>and work is underway to replace it.
>Until then my advice to get the V1 versions of the MIBs you need from the
>vendor. If they are using the SMICng compiler, these are simple to make as
>this
>is just an option you set and the compiler will produce a V1 MIB from V2
>input.
>If you absolutely cannot get a V1 version of the MIB, and must use a V2
>version,
>then you may have to edit the MIB and remove sections which are not
>currently
>supported. Tivoli NetView Support will be glad to assist you if you need
>help.
>But it is still the case that some vendors produce bad MIBs and so
>sometimes,
>when a MIB won't load, it is not our fault. In these situations, too,
>NetView
>Support will help you modify the MIB so that it will load.
>
>I hope this helps everyone's understanding here, but if not, feel free to
>ask
>further questions. I know the current situation is less than ideal, and
>all I
>can suggest until it is properly addressed is that you be patient.
>
>Thanks
>
>James Shanks
>Tivoli (NetView for UNIX and NT) L3 Support
>
>
>
>Ken Garst <Ken.Garst@KP.ORG> on 03/02/2000 02:51:22 PM
>
>Please respond to Discussion of IBM NetView and POLYCENTER Manager on
>NetView
> <NV-L@UCSBVM.UCSB.EDU>
>
>To: NV-L@UCSBVM.UCSB.EDU
>cc: (bcc: James Shanks/Tivoli Systems)
>Subject: Re: How to load and configure cisco mibs and traps?
>
>
>
>
>The key question is to ask the Tivoli/Cisco people which versions of the
>mosy
>compiler are being used. NetView is notorious for always having the oldest
>version of the mosy compiler which gives the error messages you are seeing.
>
>If you are creative, you can edit the Cisco mib .my file and correct the
>offender if you know what you are doing or simply comment out the entry if
>you
>don't. In either case, NetView can then load the mib.
>
>Regards,
>ken
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
|