04-26-2010 03:36 AM
Hi all,
We are currently monitoring approximately 50 locations each having one or more cisco devices, Catalyst 3750 and 3560.
Locations are connected via CWDM.
We would like to monitor interface errors, signal loss and power on CWDM SFP transceivers on Catalyst 3560 and Catalyst 3750 switches.
We tried to get these values (as shown using „sh interface transceiver“ command) using SNMP (through cisco-entity-sensor-mib) but we didn't get any SNMP result.
What would be the best approach to monitor these values?
What MIB or other functions/modules/features need to be used/activated on switch?
Do we need any additional piece of hardware?
Regards,
Marko
05-03-2010 11:18 PM
>> The MIB files are there just to give context and meaning to the values returned by the agent.
I knew that
Before you sent me the new MIB, i actually changed the old one myself, because "IOS will ALWAYS send 14" and "The MIB files are there just to give context and meaning to the values returned by the agent". So i figured. I changed the mib (added dBm) and expected to get dBm. But it was still 14. so i got confused and i wanted to double-check on the forum here....
I obviously didn't make the MIB changes in a proper way. But that's not important anymore, i got dBm using the MIB you attached.
Okay, the picture is much much clearer now, i have a proper MIB, i have a switch that supports the mib, i have a CWDM SFP with DOM support, and i'm getting all the optical monitoring values in my MIB browser client (or NMS).
The last thing for me to do is to use a Performance Management tool to collect these values on a regular basis i process this data for various usage (threshold, reporting, etc.).
I have to figure out the indexing mess in these MIBs (entPhysicalIndex is not accesible, etc.) to link the interface (from RFC1213-MIB-2, via ENTITY-MIB) with its SFP and (from SENSOR-MIB) to the optical monitoring measurements...
but that's another story.... if you have any comment on this last part, feel free to add something you think could help me. If not, thank you for your help, and to other guys.
Regards, Marko
05-04-2010 10:33 AM
Non-accessible indexes are nothing new. What you need to do is extract (in this case) the last element from the left-hand-side of the equal sign (so to speak). That is, take the OID returned in a GET-NEXT query, and extract the last dotted element. That will be your entPhysicalID.
05-04-2010 11:37 PM
Yes, e.g. snmpwalk returns this entPhysicalID as a last dotted element, but unfortunately, we're planning to extract and process these values using a performance tool (HP OV Performance Insight - OVPI) and that tool doesn't (I assume, i'm not 100% sure) post full OID in number format (or string format) into a database, in any column, so once we have this data in DB, there will be no OID to parse the entPhysicalID out and tie it somehow to interface (index).
I guess we'll have to pull only the entPhysicalEntry table (from ENTITY mib) and entSensorValueTable (from SENSOR mib), the OVPI will add the device name as identifier to every row so we'll know what sensors belong to device, and in another table what values belong to device, but we'll have trouble matching what values belong to what sensor (in the context of the same device). The problem will be if the device have more than 1 SFP, and every SFP has 5 sensors.
E.g.
extracted from ENTITY MIB with PerfMon tool (Device1 has 2 SFPs on Gi0/1 and Gi0/2, with 5 sensors each):
Device | entPhysicalDescr | entPhysicalClass |
Device1 | GigabitEthernet0/1 Module Temperature Sensor | sensor |
Device1 | GigabitEthernet0/1 Supply Voltage Sensor | sensor |
Device1 | GigabitEthernet0/1 Bias Current Sensor | sensor |
Device1 | GigabitEthernet0/1 Transmit Power Sensor | sensor |
Device1 | GigabitEthernet0/1 Receive Power Sensor | sensor |
Device1 | GigabitEthernet0/2 Module Temperature Sensor | sensor |
Device1 | GigabitEthernet0/2 Supply Voltage Sensor | sensor |
Device1 | GigabitEthernet0/2 Bias Current Sensor | sensor |
Device1 | GigabitEthernet0/2 Transmit Power Sensor | sensor |
Device1 | GigabitEthernet0/2 Receive Power Sensor | sensor |
extracted from SENSOR mib with PerfMon tool (2 sets with 5 values per SFP):
Device | entSensorType | entSensorScale | entSensorPrecision | entSensorValue |
Device1 | celsius | units | 1 | 483 |
Device1 | voltsDC | units | 2 | 324 |
Device1 | amperes | milli | 1 | 332 |
Device1 | dBm | units | 1 | 32 |
Device1 | dBm | units | 1 | -130 |
Device1 | celsius | units | 1 | 500 |
Device1 | voltsDC | units | 2 | 345 |
Device1 | amperes | milli | 1 | 331 |
Device1 | dBm | units | 1 | 30 |
Device1 | dBm | units | 1 | -128 |
So, with this i can't be 100% sure to match "GigabitEthernet0/1 Transmit Power Sensor" from the first table with its value in the second table, and respectively "GigabitEthernet0/2 Transmit Power Sensor" to its value.
So this is my next problem.... i don't know how to get past this. I'll have to make some tests with OVPI first, maybe something will come up...
04-12-2011 11:14 PM
Hi, could you share the perfmon configuration to get it?
thanks a lots
04-12-2011 11:43 PM
Hi,
i don't quite understand what you mean by "perfmon" configuration.
We're trying to get this info via SNMP....
marko
04-13-2011 12:00 AM
Yes, im trying this too. I see that you use perfmon tool to obtain the values and i need to know how to configure this.
Thanks a lots.
04-13-2011 12:06 AM
Yes, we're using a certain performance monitoring tool, but this is step 2 of this story.
step 1 is the SNMP indexing issue which is preventing us from using this SNMP data in our chosen perf. mon. tool.
I don't know what perf. mon tool are you using, we're planning to use this data in HP OVPI, but this data can be utilized in any other perfmon tool.
Anyway, configuring a performance tool is not a topic here. The topic is how to relate two bits of SNMP data into something that can be used by any perfmon tool.
regards
marko
04-13-2011 12:07 AM
but if you mean how did i get those tables posted earlier, these are form IReasoning MIB browser....
04-13-2011 12:11 AM
Yes, because i want to discover what values belong to each interface.
Thanks for your fast response
04-27-2012 01:29 AM
Hi Marko,
Did you ever find a solution to your problem?
I'm in the exact same situation as you and I wish to monitor a number of transceivers on a number of different platforms.
Thanks!
07-12-2012 07:25 AM
hi all i need help i try all what you write here .
i have
Vendor Name : CISCO-FINISAR
Vendor Part Number : FTRJ1319P1BTL-C
Vendor Revision : 0x42 0x20 0x20 0x20
and when i try to get info from telnet i get normal information
#sho interfaces transceiver
If device is externally calibrated, only calibrated values are printed.
++ : high alarm, + : high warning, - : low warning, -- : low alarm.
NA or N/A: not applicable, Tx: transmit, Rx: receive.
mA: milliamperes, dBm: decibels (milliwatts).
Optical Optical
Temperature Voltage Current Tx Power Rx Power
Port (Celsius) (Volts) (mA) (dBm) (dBm)
--------- ----------- ------- -------- -------- --------
Gi0/26 52.5 3.25 26.9 -5.7 -13.8
but i couldnot get from snmp can some one tell me if is some configure in device or oid i try is not thet what i need .
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide