Start a Conversation

Unsolved

This post is more than 5 years old

4377

June 13th, 2013 07:00

load MIBS into SMARTS system

Hi,

We use EMC SMARTS as the Monitoring tool in our Data Center.  Right now we are integrating few new solution into our mgmt stack. One of them is a backup tool providing MIBs to enable snmp monitoring. We would like to have this in place with SMARTS so that any backup task failures can be alerted in SMARTS console. Our monitoring stack includes In-Charge AM-PM, ESM, OI, SAM and Global console (alert dashboard). In-Charge-OI is configured as the snmp trap listener.

Few questions I have - 

How do we load specific MIBs into SMARTS system? And what other configuration required for snmp polling after MIBs are loaded?

If we need to read specific OID during snmp polling, what steps need to be taken?

Appreciate your suggestions/tips. Thanks!

30 Posts

June 13th, 2013 09:00

Hi

You can achieve this with dynamic modelling.

Following components are needed:

- A DM (dynamic modeling) model (license is required as well)

- GA_Driver

- Discovery Script

- reconfigure script

- setting script

- good know how of ASL

Therefore you need to enhance the your certification. In our environemnt we did it with the following entry:

= containment

Therefore you will need to modify ./rules/discovery/ic-oid2type.asl script. That's because the parser will not be able to load the new added certification into IP manager.

Afterwords you have to write the discovery process, wher you define which MIB information (static) you will discover to create your own instances.

The instances are base on dynamic modeling, e.g.

interface ICIM_System;

interface SC_c3gWirelessInstrumentation:ICIM_Instrumentation

"Instrumentation for Cisco 3g Wireless WAN Interface Cards for ISRs. "

{

instrument SNMP {

SC_c3gGsmCurrentServiceStatusInt="1.3.6.1.9.9.661.1.3.2.1.2"

};

readonly instrumented attribute int SC_c3gGsmCurrentServiceStatusInt "Current Service Status: 1=unknown, 2=noService, 3=normal, 4=emergencyOnly."

"OID: 1.3.6.1.9.9.661.1.3.2.1.2";

}

interface SC_c3gWireless:ICIM_LogicalElement

"Cisco 3g Wireless WAN Interace Cards for ISRs."

{

stored attribute string DeviceID "" = "NA";

stored attribute string SC_c3gImsi "Interantional Mobile Subscriber Identity (IMSI)." = "";

stored attribute string SC_c3gImei "Interantional Mobile Station Equipment Identity (IMEI)." = "";

stored attribute string SC_c3gFsn "Factory Serial Number (FSN)." = "";

internal stored attribute int SC_c3gGsmPdpProfileTypeInt "PDP (Packet Data Protocol) Address."=1;

stored attribute string SC_c3gGSMPdpProfileApn "Access Point Name (APN)." = "";

stored attribute string SC_c3gGSMPdpProfileAddr "Configured PDP (Packet Data Protocol) address." = "";

internal propagate attribute int min SC_c3gModemStatusInt "Modem Status." <= SC_c3gWirelessInstrumentation,InstrumentedBy,SC_c3gModemStatusInt;

enum c3gModemStatus_e {UNKNOWN_MODEMSTATUS=1,OFFLINE=2,ONLINE=3,LOWPOWERMODE=4};

/*

  • computed attributes to be more meaningful

*/

computed attribute c3gModemStatus_e SC_c3gModemStatus "Modem Status." =

c3gModemStatus_e(SC_c3gModemStatusInt) else UNKNOWN_MODEMSTATUS;

/*

  • thresholds defintion

*/

attribute int SC_c3gModemTemperatureThreshold "Threshold for temperature value. 0 does disable this event." = 50;

attribute int SC_c3gCurrentGsmRssiLowThreshold "Threshold for RSSI threshold. 0 does disable this event." = -100;

attribute int SC_c3gCurrentGsmRssiMediumLowThreshold "Threshold for RSSI threshold. 0 does disable this event." = -90;

/*

  • event definitions

*/

event SC_c3gServiceNotNormal "Indicates that the current status is on status noService(2) or emergencyOnly(4)." =

if (SC_IsManaged && !IsUnmanagedOrUnresponsive)

check (SC_c3gModemStatusInt== 2 || SC_c3gModemStatusInt == 4);

}

In this model you instrument all values which are dynamic (changing over time), e.g. in this model we poll the status of the this 3G GSM modem.

This model you will have to compile with the command dynmodel. There is an Dynamic Modelling guide available from EMC.

To load this new class into IP manager you will have to reload the IP domain and you have to have --dynamic option enabled for the domain.

Afterwords you need the GA_Driver which will be used during discovery and a health script which will interpret/translate the SNMP information you get during discovery.

The GA_Drvier you have to define in conf/discovery/DISCOVER_.import and looks like follows:

GA_Driver::SC_c3gWAN-Driver {

ReadsRulesFrom = GA_RuleSet::SC_3cgWAN-RS

{ fileName = "sc_info/sc_c3gwan-health.asl" }

ReadsInputFrom = SNMP_WalkFE::SC_c3gWAN-FE {

retries = 5

rootOIDs =

{

{10, ".1.3.6.1.4.1.9.9.661.1.3.1.1.1"}, #c3gIsmi

{20, ".1.3.6.1.4.1.9.9.661.1.3.1.1.2"}, #c3gImei

{30, ".1.3.6.1.4.1.9.9.661.1.3.1.1.5"}, #c3gFsn

{40, ".1.3.6.1.4.1.9.9.661.1.3.3.1.1.2"}, #c3gGsmPdpProfileType

{50, ".1.3.6.1.4.1.9.9.661.1.3.3.1.1.3"}, #c3gGsmPdpProfileAddr

{60, ".1.3.6.1.4.1.9.9.661.1.3.3.1.1.4"} #c3gGsmPdpProfileApn

}

}

waitForCompletion = TRUE

}

The asl health script will create the instances. And add it to the topology.

Further you need to modify an existing reconfigure/setting script or you can write your own reconfigure/setting script, which will be run during reconfiguring.

This reconfigure/setting script is located under rules/health. If you decide to have your own script you need your own pollinggroup created with dynamic modeling.

This script will create the needed instrumentation and accessors that the information are polled.

In the model above you are able then to create your own events.

If you need any further information, please let me know.

Regards,

Christian

23 Posts

June 13th, 2013 10:00

Hello,

If you are intended to configure Smarts to deal with the received SNMP Traps from your third party nms, you have to work with trap_mgr.conf.

This file is located under your OI Instalation (" /conf/icoi/") and it basicaly map SNMP Varbinds to Smarts Notifications.

There are several other things you can do and I recomend you to take a look at SAM OI Adapters Documentation.

NOTIF can do the work for you if you have the proper license.

Regards

Gabriel

22 Posts

June 14th, 2013 07:00

Thanks much for these details. SNMP traps from the tool now shows up in InCharge-OI and then on Global console dashboard. However looks like more changes required to adjust the alerting.

Every events (traps) sent by the tool are reported to a "single" alert (named as SNMP Trap Catchall) in InCharge-OI, and whenever a new event occurs, it overwrites previous notification alert and log those previous into Audit log of the alert. Do you know how to make this change, so that every trap is reported as a New separate alert?

Secondly, the trap collected does not show the result of a  particular oid that returns. I did some tests with the tool itself as the snmp trap collector where I received those text for same event.

As an example, this is the trap received on the 3rd party tool.

TRAP, SNMP v1, community public

iso.1.5.1.3.1.13987.1.1.1.2.4 = STRING: "ERROR" iso.2.5.1.3.1.13987.1.1.1.2.5 = STRING: "Task failed - API error."   

In case of InCharge-OI, I don't see the above results (i.e Task failed -API error is not shown). Can this be fixed through trap mgmt  configuration file?

Sorry, I am not much familiar on snmp concepts but could you please share your thoughts on these issues.

Thanks again! 

23 Posts

June 14th, 2013 08:00

Hello,

You will have to google some information about SNMP. There are lots of information available.

What you have to understand to configure Smarts is how information is encapsulated under SNMP Trap package.

Aparently, your events are matching the last rule of trap_mgr.conf file. (This has to be the "catch all" rule).

You need to create new rules on this file.

First of all, add this line to the rule: "LogFile:      somefile.log"

This will log every trap received under smarts receiver log folder.

You will be able to create new rules on trap_mrg.conf from there.

5 Practitioner

 • 

274.2K Posts

July 24th, 2013 02:00

Hi vtrack,

If you send me the MIB file I can send you back the trap_mgr.conf entries for the configured traps.  After that I am guessing that you will need to do some changes to the entries and I am also guessing that you will need to do some changes through an ASL hook since I am again guessing that you will need to process some of the varbinds with logic.  My email is ray.pramuka@emc.com

Ray

22 Posts

July 24th, 2013 23:00

Thanks Ray for offering your helps here. We got this working few days back with NOTIF and adding specific varbind values added to Event fields making each event unique. Thanks again.

No Events found!

Top