Object IDs shown in Hardware Sentry KM do not Match with Other Management Consoles
KB1061 - Mar 25, 2011
Description: Why object IDs shown in Hardware Sentry KM do not match with other management consoles.
Additional Keywords: Hardware Sentry KM for PATROL
The numeric object IDs displayed in the PATROL Console (or in the BMC ProactiveNet Performance Management console) do not match with the object naming in the other management consoles.
Example: "Volume: 31 (1.0 GB)"
The numeric instance ID generated when a hardware failure is detected and displayed in the PATROL event does not match with the object naming in the other management consoles.
The problem typically happens when monitoring a disk array through an SMI-S compliant agent (notably EMC, IBM DS 6000/8000 and Hitachi disk arrays).
No specific symptom: the object IDs shown by Hardware Sentry KM look normal until they are specifically looked for in another management interface.
Object IDs are always very long strings when monitoring a system through WBEM (EMC, IBM and Hitachi implemented the SMI-S instrumentation standard, which itself is a WBEM/CIM derivative), which is not very handy to the user and cannot reasonably be displayed in the Console (and BEM, etc.) treeview.
Example of such ID:
(This is the ID of a disk in an EMC Clariion disk array)
As a workaround, Hardware Sentry KM assigns numeric IDs that are easier to read. That’s the number that you read in the icon name "Volume: 1 (1.0 GB)". This number is purely artificial and its sole purpose is to let you discriminate numerous icons in the treeview of the PATROL Console. When an alert is triggered on this object, you need to look at the internal "Device ID" shown in the infobox, which will display the original (and lengthy) ID and therefore should contain all of the information to clearly identify the faulty device. Additionally, when possible, Hardware Sentry KM shows additional “Identification Information” in the infobox.
In ordinary cases, Hardware Sentry KM uses the real object ID in combination with the hostname and name of the connector used to monitor the system to create an ID of the PATROL class instance. However, the total length of the PATROL object path (/MS_HW_PHYSICALDISK/…./Status) has a hard limit and cannot exceed 126 characters, which means that the instance ID itself cannot exceed 100 characters in most cases. This is not sufficient to display the ID of the monitored object as reported by a WBEM-instrumented system.
As a workaround to that limitation, Hardware Sentry KM assigns another numeric ID to each object that has a device ID too long to be included in the PATROL Object ID. This is the ID you can read in the instance ID and PATROL Object Path "/MS_HW_LOGICALDISK/1001/Status". Again, this number is purely artificial and its sole purpose is to circumvent an internal PATROL Agent limitation regarding object IDs. When an alert is triggered on this object, the internal instance ID should not be taken into account. Instead, you will need to look at the internal "Device ID" shown in the Infobox, as explained previously.
The table below summarizes the various IDs that are assigned to an object.
|Device ID||Real device ID as retrieved from the instrumentation interface.
This is the ID that should be looked at when trying to locate a faulty device.
Hardware Health Report
|PATROL instance ID||Artificial number assigned by Hardware Sentry KM for device IDs that exceed the PATROL’s ID length limit.||/MS_HW_PHYSICALDISK/1173/Status||Internal PATROL object ID and path
PATROL events (origin)
Thresholds configuration (Event Management KM, PCM)
|PATROL instance name||Artificial number assigned by Hardware Sentry KM to discriminate several instances of the same class in the PATROL Console||Volume: 31 (1.0 GB)||PATROL Console treeview
BMC ProactiveNet (BPPM) console