Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

2109

March 22nd, 2017 03:00

Question on SRM's snmp oid customize

Hi,

     I reconfigured some MIB OID from SRM "MIB Browser" and "SNMP Collector" for no SolutionPackage storage device, like CPU Memory Disk IOPS.

     Almost all the configurations used default ones generated by "MIB Browser", now we can see the metrics data for controller's CPU Memory, but no data for LUNs and nodes' IOPS.

     There are 2 controllers there, we can see 2 lines in report. These are 2 Nodes there, totally no data. There are many LUNs that may change by time, also no data.

     May I know how to configure collection for IOPS and LUNs?

141 Posts

March 23rd, 2017 06:00

Hi Jianping.Wang,

You should check the collector-manager logs for the snmp-collector that you are using; by default it is Generic-SNMP. You can also increase the logging level by editing the .level value in the logging.properties file, you can increase it to FINE, FINER or FINEST, please note that FINER and FINEST are very verbose and most snmp polling issues should identifiable with the default INFO level.

Unfortunately i cannot provide a definite solution on how to poll the IOPS and LUN data trough snmp as this depends on the SNMP agent and the mib that it is being used.

You can open a service request for assistance with basic functionality however support is limited on custom mask creation. Another option would be to contact professional services to have them assist or create the mask for you.

hth

Regards,

Isaka

March 23rd, 2017 04:00

Hi,

      Let me resolve this in another way, when I cannot see the polling result data, which log should I check to know the root cause? And please also guide me how to change log debug level if needed, thanks.

March 25th, 2017 02:00

Hi Isaka,

     Thank you for guide, I can see more logs now, but no message related to specific metrics. I can see "java.lang.ClassNotFoundException" in log, and I guess wrong parttype will cause problem, is there any document to describe what kind metric maps to which parttype?

141 Posts

March 28th, 2017 05:00

Hi Jianping.Wang,

I don't think this caused by the parttype, as it is a only property of the metric, as a rule of thumbs up we want the parttype value to be meaningful (interface, port, processor, etc...) for reporting purposes but could be whatever you like (with some exceptions when using index).

The stack trace should include a "caused by" statement that would provide more information about the missing class.

regards

Isaka

March 30th, 2017 00:00

Hi Isaka,

     Thanks for guide.

     May I know where and how to check stack trace? I cannot find it from manuals.


141 Posts

March 30th, 2017 10:00

Hi Jianping.Wang,

Stack trace refers to the full error message.

Regards,

Isaka

April 6th, 2017 18:00

Hi Isaka,

     Thanks for help so long,

No Events found!

Top