Cisco-CatOS-MIB

[torstetfk]

Hello,

it would be nice to have the same naming possibility for graphing at Cisco CatOS (I mean the description when you enter "show port stat" on the Catalyst) as we have for Cisco IOS (IfAlias of IF-MIB).

At IfAlias we have the interface description which ist manually entered at configuration time on the router. I use the IfName instead of IfIndex to get a better headline in the graph.
With CatOS I get only the portname (example: 3/12). So I must rename each graph to be more speaky.

I suggest to make a "Host Template" for Cisco Catalysts that checks for the right MIB to get the better naming for the graphs.

Greetings
Torsten

[bulek]

You may try to use the following command on your switch :

set snmp ifalias [ifIndex] [interface_description]


- bulek

[torsten]

You may try to use the following command on your switch :

set snmp ifalias [ifIndex] [interface_description]


- bulek

Hi,
that's an alternative. But then I use the IfIndex which changes at next reboot. I also have to duoble work because I set the port names with "set port name xxxxxxx".
Do you know the MIB to get this names per SNMP? I could hack the php-scripts to fetch this MIBs.

Greetings
Torsten

[bulek]

That's right (the double work I mean). Unfortunatelly I don't have any CatOS based switch and don't know the MIB with these values. You may try to do snmpwalk with OID .1 on one of your switches. Then look for the names you entered in config.

- bulek

[dbee]

Hi,

try this oid:.1.3.6.1.4.1.9.5.1.4.1.1.4 for printing the interface_description.

hope this helps

dbee

[torsten]

Hi,

try this oid:.1.3.6.1.4.1.9.5.1.4.1.1.4 for printing the interface_description.

hope this helps

dbee

OK this helps - but then you must change every MIB in the interface.xml file to get the right index for inserting it into the mysql db.
If you don't change it the names and interfaceses don't match each other.
I think the MIB for receiving InOctets is also not ok if you use the old MIB and the new one.
Changeing to the MIB-tree .1.3.6.1.4.1.9.5.1.4.... aggregated ports are not visible in cacti.

So I must wait for changing CatOS to IOS....

See you
Torsten

[Guest]

yes thats right, i´ve printed out the other values of this mib so you can
build your own xml file for the catalyst switches or you can wait until next week > then i´ve got a little bit more time and can test it.

.1.3.6.1.4.1.9.5.1.4.1.1.

portModuleIndex (1)
portIndex (2)
portCrossIndex (3)
portName (4)object details
portType (5)
portOperStatus (6)
portCrossGroupIndex (7)
portAdditionalStatus (8)
portAdminSpeed (9)
portDuplex (10)
portIfIndex (11)
portSpantreeFastStart (12)
portAdminRxFlowControl (13)
portOperRxFlowControl (14)
portAdminTxFlowControl (15)
portOperTxFlowControl (16)
portMacControlTransmitFrames (17)
portMacControlReceiveFrames (18)
portMacControlPauseTransmitFrames (19)
portMacControlPauseReceiveFrames (20)
portMacControlUnknownProtocolFrames (21)
portLinkFaultStatus (22)
portAdditionalOperStatus (23)
portInlinePowerDetect (24)
portEntPhysicalIndex (25)

[BitFlipper]

I am not sure if what I've learned or seen will help anyone or not. I didn't see (or maybe I misunderstood) anyone describing the behavior I am seeing.

The short version . . . the information is all there. Just not sure how to walk around the 'nested' indexes.

Using a Cisco 6509 runing CatOS v7.6(3)

Here are the OID's as I've found them with brief descriptions:

mib-2.interfaces.ifTable.ifEntry.ifIndex
.1.3.6.1.2.1.2.2.1.1.<index>
Interface index found in the default interface.xml. Contains the same value as <index>

mib-2.interfaces.ifTable.ifEntry.ifInOctets
.1.3.6.1.2.1.2.2.1.10.<index>
mib-2.interfaces.ifTable.ifEntry.ifOutOctets
.1.3.6.1.2.1.2.2.1.16.<index>
Input / Output octets for a given Index

private.enterprises.cisco.workgroup.ciscoStackMIB.portGrp.portTable.portEntry.portIfIndex
.1.3.6.1.4.1.9.5.1.4.1.1.11.<blade>.<port>
Where <blade> is the blade number on your 6500 chassis and <port> is the port number on that particular blade. So .1.3.6.1.4.1.9.5.1.4.1.1.11.7.13 would contain the index value of port 13 on blade 7. This value (from what I've seen) does cross reference into ifIndex (.1.3.6.1.2.1.2.2.1.1)


private.enterprises.cisco.workgroup.ciscoStackMIB.portGrp.portTable.portEntry.portName
.1.3.6.1.4.1.9.5.1.4.1.1.4<blade>.<port>
Where <blade>.<port> are the same as above. Contains the port name (description) everyone seeks ie. "my uber server"


mib-2.ifMIB.ifMIBObjects.ifXTable.ifXEntry.ifName
.1.3.6.1.2.1.31.1.1.1.1.<index>
Contains the port name often used ie. "7/13"

The piece that I think gets confusing is the <blade>.<port> conbination. It mixes things up where portnames are not associated with the right index/interface.

I have not created a single .xml file for a single blade yet and not even sure if that is the right direction. Although, segmenting the blades would be a bonus. Waiting for 100 images to generate can get tedious.

I did notice this 'issue' is a topic for some people per a reference made in this week's beta post from TheWitness.

Hope this helps.

[beeudoublez]

I've searched around the forums and I can't seem to find anything new for this Cisco Catalyst parsing, has anyone had any success?

[TheWitness]

Need to write a new Interfaces.xml to make it work called something like "CatInterfaces.xml". Then utilize the OID_INDEX_PARSE to peel off the last two OID components from the CatOS OID's to get "port.slot" information. The problem then becomes that there is then a mapping from the "port.slot" OID reference to the ifIndex MIB.

It kind of sucks. It's been kind of a long term assignment for me but it's off the bottom of the list. I could do it from home, but someone must place a CatOS switch on the internet for me to query.

TheWitness

[beeudoublez]

Okay, I honestly haven't stared too hard for a solution, I understand the problem with the way CatOS stores it.

So am i correct in that Cacti has the capability to do this blade/port mapping and someone just needs to write the XML file? Or does Cacti need to be modified? I'm still wrapping my head around this so I can't really see where the gap is....

I looked into throwing a CatOS on the internet for ya, but it didn't get too far

[TheWitness]

You know, I was thinking of this. I think it needs to be written like a script. However, then all the benefits of multiple OID get's will not be achieved.

I need to think about this some more. Stay tuned.

TheWitness

[BitFlipper]

TheWitness,

Would it help to have a flat text file to grind through as input for script testing? I realize it is not quite the same as the actual equipment but maybe would help get the basic logic down.

[TheWitness]

I discussed this issue with the lead developer yesterday and a new data collection method needs to be implemented. Once I get the wrinkles out of Cactid 0.8.6f-1, this will be my next priority. Although, people are clammering for the End Device Tracking Add On.

Larry

[rony]

The new method is going to have to wait until 0.9.0. Because it's a major change in the way input methods work, and the potiential to break so many things in 0.8.6.