Missing metrics in UMP USM portlet for Cisco 3850 and 4510 model switches monitored with the snmptoolkit probe

Document ID : KB000004371
Last Modified Date : 14/02/2018
Show Technical Document Details
Issue:

The snmptoolkit is configured to monitor Cisco 3850 and 4510 model switches because the cisco_monitor probe is not designed to monitor these models.  CPU and Memory QoS metrics are being collected successfully, as can be seen using the DrNimbus tool as well as using the UMP Performance Report Designer, but when viewing the devices from the UMP USM portlet metrics tab, no network metrics can be found for the devices.

Environment:
- UIM Server through GA release 8.47- snmptoolkit versions through GA release 1.39
Cause:

The checkpoints in the snmptoolkit.dta file for the CPU and memory metrics were configured with the following key values:

CIType = 99
CIMetric = 500

The snmptoolkit states the following for these fields:

CIType/CIMetric
Probe-specific and checkpoint-specific values. The probe automatically sets these values. Leave these elements blank.

By specifying a value, the defaults that the probe uses are overwritten to use the CIType specified in the snmptoolkit.dta file for the checkpoint.  Currently a CI type of 99 does not exist in the CM_CONFIGURATION_METRIC_DEFINITION table, so the search for metrics for the devices using these checkpoints were never associated with the devices because a matching CI type was never made.

Resolution:

The following steps will correct this:

 

1. Deactivate the snmptoolkit probe

2. Manually modify the snmptoolkit.dta file and remove all CIType and CIMetric settings from all QoS checkpoints

3. Manually modify the snmptoolkit.cfg file and remove all CIType and CIMetric settings from all profile QoS checkpoints

4. Reset the ci_metrid_id column in the S_QOS_DATA table to NULL for the snmptoolkit probe (this is necessary because you should get a new set of metric IDs now that the ci_type will be different).  SQL Query to do this is as follows:

    UPDATE S_QOS_DATA SET ci_metric_id = NULL WHERE probe = 'snmptoolkit';

5. Restart the data engine so that it will repopulate the S_QOS_DATA table with the new ci_metric_id once you restart the snmptoolkit probe

6. Clear the niscache directory on the robot where you have the snmptoolkit probe deployed using the _nis_cache_clean command from the controller probe utility in Expert Mode on the robot

    a. Select the controller probe on the robot where the snmptoolkit probe is deployed from Infrastructure Manger (IM) and enter <ctrl>p or select the probe from Admin Console, then select the Probe Utility from the probe’s drop down menu list

    b. From the IM Probe Utility dialogue, select the Options button (second from the right with the yellow gear), select the ‘Expert mode (no confirmation required)’ option, then click OK

    c. Select the _nis_cache_clean command from the Probe commandset drop down list

    d. Select the play button (green arrow) to execute the command

7. Restart the snmptoolkit probe