On Tue, 2005-04-12 at 09:43, Frithiof Jensen (AH/LMD) wrote: > Just to make sure that I understand things correctly
Always a good idea... > - after all I am fairly new to this, so please bear with me. Of course. > The Way It Should Work (tm): > > The IP-FORWARDING-MIB.txt that comes with net-snmp 5.2.1 was > used by mib2c to compile the skeleton code ip-forwarding-mib > Implementation, Agreed. > part of which is the inetCidrRouteTable. I'll come back to that bit in a moment. > The tools use the same IP-FORWARDING-MIB.txt to represent > "the bucket of bits" recieved from the Agent. Because the > same MIB was used when building the Agent and the Tools, > everybody is happy. Yup - that's the way Things Ought To Work. > The Way It it may be Broken: > > The IP-FORWARDING-MIB that comes with netsnmp 5.2.1 is outdated, > but due to an unfortunate accident, the latest IP-FORWARDING-MIB > was used by mib2c to generate the ip-forwarding-mib implementation. > Thus the Agent is correct, but the Tools are confused. Nope - that's not quite correct. The IP-FORWARD-MIB that comes with net-snmp 5.2.1 is the most recent released version. This MIB was used to produce the agent implementation modules, including mibII/ipCidrRouteTable (but *not* inetCidrRouteTable, which isn't part of this version of the MIB). The (default) agent and the tools are consistent. > The IP-FORWARDING-MIB that comes with netsnmp 5.2.1 *is* the latest, > but the one used to generate the ip-forwarding-mib implementation is > an obsolete one. That's not quite correct either. The IP-FORWARD-MIB that comes with net-snmp 5.2.1 is indeed the latest released version. The one used to generate the ip-forward-mib implementation is a newer, *unreleased* version. Both this MIB, and the implementation of it (mibgroup/ip-forward-mib) are therefore "experimental", and are not included in a default build. (Though as Wes points out, the MIB is sitting on the RFC queue, and is unlikely to change significantly when it does finally appear) Certainly, when the ip-forward-mib (and equivalent) implementations become accepted as part of the default build, then we'll definitely need to ship the corresponding MIB files as standard. But we're not at that stage just yet. The ip-forward-mib implementation is still very new, and is included so that we can get some feedback about how well it works (and hopefully encourage people to port it to other systems). That's my understanding, anyway - Robert may have a different take on things. I seem to remember a brief discussion about which version of the IP-FORWARD-MIB we should ship - the current one (as used by the agent on most architectures, and in the default build under Linux), or the newer version (as implemented in the experimental build under Linux only). The consensus was that the MIB file should match the more usual configuration - that it would simply cause more confusion if we shipped the unreleased version (most of which was irrelevant most of the time anyway). That may have been the wrong decision, but it made sense at the time. And I'm happy to stand by it. > Again, the agent will fetch proper information but it will drop it > into obsolete/wrong OID's The agent should *never* report information in the "wrong" OIDs. MIB revisions are allowed to mark particular objects as being deprecated or obsolete (or even drop them altogether?). But it's not valid to change the meaning of a given OID. Once it's been defined, that's that - it's fixed. (Well, it might be extended, so that the old definition may end up being incomplete, but it should still be valid as far as it goes). I haven't checked Robert's code to see how it handles "obsolete" OIDs (e.g. the ipCidrRouteTable). But judging by past practise, I'd expect the agent to continue supporting them too - reporting the same information in various guises. We still implement the atTable, despite it having been replaced by the ipNetToMedia table, more than 14 years ago! Dave ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Net-snmp-coders mailing list Net-snmp-coders@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
