[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference azur::mcc

Title:DECmcc user notes file. Does not replace IPMT.
Notice:Use IPMT for problems. Newsletter location in note 6187
Moderator:TAEC::BEROUD
Created:Mon Aug 21 1989
Last Modified:Wed Jun 04 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:6497
Total number of notes:27359

5330.0. "SNMP enterprisespecific targetting" by ADO75A::BOUCHER () Mon Jul 12 1993 20:32

    Greetings.
    
    I have a customer who has a large Extended LAN with Translan Bridges
    running SNMP as the management protocol.  In order to reduce network
    overheads he wants to utilise enterprisespecific traps from a central
    bridge in order to target a link failure (If interface 4 goes down this
    equates to Site XX being unreachable).
    
    Q:  IS this possible ?  Assume that all Translan's are in the same
    domain, so targetting should not be a problem.
    
    At the moment the enterprisespecific trap only indicates that an
    interface has gone down, but the trap comes from the SNMP entity, not
    the SNMP X interface child entity.
    
    Can it be done ?
    
    Thanks in advance.
    
    
    Reece Boucher
    Adelaide, Australia
T.RTitleUserPersonal
Name
DateLines
5330.1have to use alarms and data collectorCHEEKO::DITMARSPeteTue Jul 13 1993 14:1231
    Hi,
    
    Using the built-in targeting, retargeting a notification to a different
    entity can only be done if you can derive the correct target entity
    based on either the entity that generates the event or on the event
    itself.  You do not get the capability to do anything complex like look
    at the argument of an event, etc..
    
    It sounds like in your case the event comes from the parent
    entity and not the interface itself, and the same event arrives for
    each different interface, so you'll have to use an alarm rule and the
    data collector to do what you want.
    
    I understand that the interface is specified somewhere in the details
    of the event.  In this case, you write an alarm rule which listens for
    the event and runs a procedure which parses the event report to
    determine which interface the event applies to.  The procedure then
    uses the data collector to generate an event, and specifies that the
    target entity for that event is the correct entity (site XX, etc.). 
    You will need to set up a data collector entity in the correct domain,
    and have a notify request running for that domain that listens for
    events from the data collector.  Finally, you may want to mask the
    notification that will arrive whenever the rule fires by using filters
    from the notification PM.
    
    There are lots of examples in notes in here and in the MCC_TOOLS
    conference on putting the data collector to work.
    
    regards,
    
    Pete
5330.2getting the interface (ifIndex) takes some workCSOADM::ROTHLight fuse and retire quicklyTue Jul 13 1993 15:3060
It's a general hassle to decode these kinds of traps. Cisco and Vitalink
boxes do a generic 'link down' and 'link up' type of trap. If you write
an alarm rule that  starts a batch job, these are the kinds of items
available to your batch procedure in P1 - P8:

  P1 = "Domain LOCAL_NS:.soc Rule soc_SNMP_LINKUP "
  P2 = "A link on a Vitalink or Cisco box has come back up"
  P3 = "SNMP_LNKUP"
  P4 = "(OCCURS (SNMP * linkUp))"
  P5 = "13-JUL-1993 13:35:34.79"
  P6 = "The last event detected: SNMP LOCAL_NS:.bbsoc3 linkUp  
                13-JUL-1993 13:35:33.88"
  P7 = ""
  P8 = "SYS$SCRATCH:MCC_ALARMS_DATA_13353479.DAT"

If you look at the values, you see that no interface information is
passed in any of the params. To get the interface number (which is what
you need to be able to figure out which link went down) you must
open and read through the file named in P8. It looks somthing like this
(this sample not from same trap as above):

RULE: Domain LOCAL_NS:.soc Rule soc_SNMP_LINKDOWN 

MANAGED_OBJECT: SNMP health-ether 

DESCRIPTION: A link on a Vitalink or Cisco box went down

CATEGORY: SNMP_LNKDN

EXPRESSION: (OCCURS (SNMP * linkDown), in domain soc)

TIME: 1-APR-1993 04:38:50.31

EVIDENCE: The last event detected: SNMP health-ether linkDown
1-APR-1993 04:36:42.59

EVENT_ARGUMENTS: Event: linkDown A linkDown trap was received:
enterprise = "1.3.6.1.4.1.9.1.5"  agent-addr = 156.63.224.1  generic-trap
= linkDown  specific-trap = 0  time-stamp = 112786675  ifIndex = 2     (

PARAMETER: _DKA700:[MCC]MCC_ALARMS_LOG_EXCEPTION.LOG

DOMAIN: Domain LOCAL_NS:.soc 

SEVERITY:     Critical 

See the item "EVENT_ARGUMENTS:"? The portion of the line "ifIndex = 2"
is what is needed to figure out the link that caused the trap. As you
can guess, this takes some overhead to finally come up with the
number of the interface that caused the trap.

If anyone else has a better way to get the interface number from a trap
I'd appreciate knowing how.

Suggestion to MCC developers: Pass the "ifIndex = n" information
somewhere in one of the P1 - P8 parameters so that it is not necessary to
go through the hassle of opening/searching the MCC_ALARMS_DATA_xxxxxx.DAT
file.

Lee Roth