I had similar questions and issues with respect to interface aliases and portsec pollers on Cisco 3560 switches. I found this to be immensely helpful (slapped myself on the head when I realized no SQL queries were required for what I was doing). Put your pollers in a group together as described here and their relationship can be maintained on your web page:
Thanks for the document mate...
It hasn't completely done what I wanted it to do but it is OK for now.
Basically there are 3 MIB tables that Cisco use to monitor the multiple CPUs in a 6500. You have a MIB that holds the actual CPU utilisation value, then you have a MIOB holding a hardware index, then you have a MIB that links the hardware index to an actual freindly description of the CPU you are monitoring.
What I have had to do is to poll all three MIBs and put them all in the same table. The problem is that when I graph the CPU utilisation values, the legend of each of the values is the hardware index rather than the freindly description. I cannot find a way of substituting the labels on the graph I have produced with the nice freindly descriptions....
Have you come accross that problem before?
Thanks so far!
Have you used the Universal Device Poller Map Values --> Enumeration Map in the advanced options? Using it here for converting Cisco power supply values 1-5 so that they become Normal State, warning/non-critical, critical, shutdown, not present (basically the way the MIB is documented). When you say "friendly descriptions", I believe what I've described above is a use case. Just a hunch based on what I think I understand about your question. Translating nerdy MIB-speak into English...or whatever you type in for your enumeration.
I haven't seen helpful documentation in manuals about how to use enumeration (in case moderators are reading--request for documentation to be improved). Otherwise I'd refer you to the relevant documentation.
Hi Donald, again thansk for the reply.
I understand the enumeration and I actually use it on one of the CPU MIBs... I'll explain in more detail how I have set this up then maybe yourself or another person can recommend how to resolve my problem...
I am polling 3 different MIBs.
1. cpmCPUTotal5minRev - this holds the actual % value of all the different CPUs in the 6500 chassis. Currently 7 in ours.
2. cpmCPUTotalPhysicalIndex - This MIB provides a hardware index linking the values from the cpmCPUTotal5minRev MIB to an actual CPU index in the chassis
3. entPhysicalDescr - This MIB links the indexes from cpmCPUTotalPhysicalIndex to freindly descriptions like "Route Processor Module 5"
So, I have created a table which includes all 3 MIBs. However, every column in the table are just numbers, no descriptions from the entPhysicalDescr MIB. So I changed the poller for the entPhysicalDescr MIB so that the Labels appear as the actual freidnly descriptions instead of just numbers.
However, in doing this, all it did was use the first 7 descriptions in the whole MIB, it did not actually use the correct description for the relevant CPU I was polling.
So, as a workaround, I manually found out which Hardware Index from cpmCPUTotalPhysicalIndex maps to which freindly description and then used the enumeration method to get some freindly descriptions in the table. So at least now I have currently values next to a freindly description...
NOW..... i am trying to graph the useage historicaly using these values. I am struggling to get the freindly descriptions to be the labels of the different lines in the graph. The only thing I can use for the Labels is the hardware index values from the MIB cpmCPUTotalPhysicalIndex.
So now when we see a cpu spike on the graph, rather than knowing exactly which CPU it is, we just have to quickly cross reference the Hardware index to the freindly description.
Not ideal, but its working and it is far more acurate than the default CPU monitoring that Solarwinds provides out of the box.
After all that... any ideas? lol