[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

4474.0. "ipReachability again and alarm setup" by TLSE01::SELLES (Pierre-Jean - Toulouse -France) Mon Feb 01 1993 10:15


	following note 3112 about testing ipReachability of an SNMP agent,
i'd like to know  :
	
	 in  a customer situation , we want to be sure when an alarm is fired 
testing  the reachability of an snmp agent ( in this case , CISCO 
routers and CHIPcom hubs ) that this agent is down  ; the problem is , with the
 basic alarm (  comparison rule or change_of rule ) , the alarm fires even if
 the agent is up and running  ;
	as we are in a "test of implementation of the network" phase , the 
customer is very concerned about this .

	So my questions are :
	
	1 - does decmcc 1.3 implements the SNMP TEST using 5 pings ? ( the 
default for unix ping )  --> so i could tell the customer its problem will 
be fixed in time 

	2 - how do i implement this using decmcc 1.2.3 ?
	--> so i could have my customer wait for all the goodies of the 1.3 
and have him sign the "network test" sheet ;-)
 
	 it seems that it should be possible with a occurs_N_times rule but 
the way to do it does'nt seem trivial to me ( the example in the Alarms 
manual p5-14 is not that obvious )

	thanks for your help 


	Pierre-jean
T.RTitleUserPersonal
Name		DateLines
4474.1MOLAR::YAHEY::BOSEMon Feb 01 1993 16:3213
	DECmcc SNMP AM V1.3 does implement ICMP retries. It is available
	as a MOM attribute :

	MCC> SET MCC 0 TCPIP_AM ICMP Retries = 5

	V1.3 also supplies an ip reachability poller which generates 
	notification events on the reachability status of SNMP entities.
	This functionality allows users to automatically find out the
	reachability status of SNMP entities without having to write
	alarm rules, thereby saving valuable system resources.

	Rahul.
4474.2TLSE01::SELLESPierre-Jean - Toulouse -FranceMon Feb 08 1993 13:5816

	thanks Rahul for your good news .

	waiting for the 1.3 SSB , we succeeded in implementing the 
Occurs_N_TIMES rule over the Ipreachability rule to detect "real" problems 
	
	by the way where is the doc for the IP poller ?

	thanks for your help

	regards  PJ selles

PS : by the way , i've send you a mail to MOLAR::YAHEY::BOSE about a 
problem we have with a Lantern monitor thru SNMP ( decmcc 1.2.3 )
let me know if you have it as i had a problem with nmail
4474.3MOLAR::YAHEY::BOSETue Feb 09 1993 08:4514
>>	by the way where is the doc for the IP poller ?

	The documentation for the poller can be found in the SNMP AM
	Use book.

>>PS : by the way , i've send you a mail to MOLAR::YAHEY::BOSE about a 
>>problem we have with a Lantern monitor thru SNMP ( decmcc 1.2.3 )
>>let me know if you have it as i had a problem with nmail

	As a matter of fact, I never received the mail. You can try
	resending it to molar::bose.

	Rahul.
4474.4ICMP retries nomore than 3, still reach probs.MLNCSC::LASAGNANO!, le scarpe no !!Tue May 18 1993 13:1550
    Rahul, you said in .1  that SNMP AM 1.3 does implement ICMP retries,
    right, but when i tried the command you suggested:
    

MCC> SET MCC 0 TCPIP_AM ICMP RETRIES = 5

MCC 0 TCPIP_AM 
AT 18-MAY-1993 17:31:17 Characteristics

Problem(s) modifying attribute(s).
                           ICMP Retries = -- Invalid Value


MCC> SHOW MCC 0 TCPIP_AM ALL CHAR

MCC 0 TCPIP_AM 
AT 18-MAY-1993 17:31:36 Characteristics

Examination of attributes shows:
                      Component Version = V1.3.0
               Component Identification = "DECmcc TCP/IP SNMP AM"
                            UDP Timeout = 5
                            UDP Retries = 2
                           ICMP Timeout = 5
                           ICMP Retries = 2
               Mib Extensions Available = ( "rmon",
                                            "EXP_RMON",
                                            "cisco" )


    I just can set the icmp retry to 3 at maximum then i got that error,
    note that this means that i got two nodes firing critical for
    IPreachability down while they're up. 
    
    Doing manually the test or the UCX ping it works fine.
    
    The Ip poller works fine, but since the customer wants the objects
    to be polled (IP reachability and conseguently objects colored) when he 
    starts the IMPM, i had to set up a set of rule (snmp * ipreachability = up) 
    and others with down.
    
    Another thing that it seems strange to me is that when i'm starting the
    IP poller it asks if i want to change the retries, i change it to 5 and
    the command procedure accept the value without error, What does it mean ??
    
    The customer has a vaxstation 4060 with 48mb, VMS 5.5-2,motif 1.1
    ,ucx 2.0b, BMS 1.3(SSB)
    
    Ciao,
    				Andrea Lasagna.
4474.5MOLAR::YAHEY::BOSETue May 18 1993 17:4911
	The maximum ICMP Retries you can set for the SNMP AM is 3 (this,
	amazingly, is actually documented in the Use book :-)). The IP 
	Poller, on the other hand, is a separate application and accepts
	larger retry values.

	You may try increasing the ICMP timeout value instead of the retries.
	If the latency for those particular nodes is too high, then increasing
	the number of retries will not help.

	Rahul.