does DataColl use 18.104.22.168.22.214.171.124.2.1.19 to determine the duplex status of an interface or does it use proprietary mibs ? For some ports I can see the value '4' (integ) in ANI db which is not a defined value for this OID.
What are the allowed values for this item in the DB? Are there consistent integers across all device types so I can map the value to a string ( 1 is unknown, 2 is half-duplex, 4 ?? perhaps full-duplex) or are they device dependent (I hope not, but they are stored in device specific port tables..)
Solved! Go to Solution.
Since we do not support direct access to the ANI database, we do not guarnatee the data values will be easily consumable by users. Duplex is one such thing. The values are normalized internally since different devices support different methods for obtaining duplex. A value of 1 corresponds to unknown duplex, a value of 2 corresponds to full duplex, a value of 3 corresponds to half duplex, and a value of 4 corresponds to a default setting.
I experienced well that the data is really not easily consumable and you can get some indigestion... but I think I am through the worst part and I am looking forward to be completely recovered ;-)
It is just that RMEs` 'Detailed Device Report' is missing some basic Information like device_VTPdomain, interface_VLANname, interface_VLANid, interface_PortDuplex
So to get a real 'Detailed Device Report' I have got to pull this out of ani db and merge it with the rme report.
I think it is worth to be discussed at Cisco to provide a dump (complete device dump) of ani db in the matter it can be done with rme ( or to complete this missing details in the rme report)
One last thing, is it correct that if 4 corresponds to a default setting it is used for interfaces which duplex mode cannot be changed (e.g a gig-if always has full-duplex) but it is not necessarily full-duplex but also can be half-duplex ( but I?ve got no example for this...)
I did a quick check and found portDuplex = 4 on several cat3550 devices and one WS-C6006 running CatOS:
cisco.1.431 12.1(13)EA1c and 12.1(22)EA2 gi0/x
cisco.1.366 12.1(13)EA1c and 12.1(22)EA2 fa0/x
cisco.1.367 12.1(13)EA1c and 12.1(11)EA fa0/x and gi0/x
cisco.1.368 Version 12.1(13)EA1c gi0/x
cisco.5.38 CatOS 6.3(5) with ifAdmin: DisabledOrDecommissioned
AND ifOper: StoppedOrSuspended
on ports of a WS-X6348-RJ-45 module
Actually, switches that support the STACK-MIB are special. In this case, default actually means the port is set to auto mode for speed and duplex. This is why we don't duplish these things.
hello Joe, I have got to come back to this once more...
as you explained, the values for duplex settings are normalized to '1','2', '3', '4' but could have different meanings for different device types. Currently I assume the interpretation of these values could be bound to the ClassListDbId of a device (or port). Could you confirm this and if this is truth must these values be interpreted along with the current ifAdminStatus and ifOperStatus of a port or can they be interpreted independently (according to their ClassListDbId)?
Yes, class of port does come into play here (which is why the CISCO-STACK-MIB devices behave differently). ANI does not consider interface status when determining duplex. It uses what the device tells it about a given interface.