[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

4563.0. "IP RECHABILITY fails" by ZUR01::SCHNEIDERR () Thu Feb 18 1993 03:17

Hello,

We run DECmcc V1.2 on ULTRIX on a DECstation 5200 with 96 MB of memory. We have 
about 150 rules that checks IP RECHABILITY (in fact PING, ICMP).
We poll all 5 Minutes. We have for each entity an own rule, no wildcards are
used.

Now we get a lot of alarms "NOT RECHABLE". If we do a TEST right after the
alarm, MCC sees the SNMP node rechable. We had a look to the network with the 
analyser an saw, that the ping message from the MCC is answered by the SNMP
station within 350 ms but the alarm is fired. So the ping answer must be loosed
somwere in the DECmcc.

Does someone has a answer to that or a recomondation?

I thought about to enable the alarms with a timedelay between or to do wildcard 
rules for each domain.

We run MSU parallel to MCC. MSU works o.k.



Roland

T.R	Title	User	Personal Name	Date	Lines
4563.1	re .-1	ANOVAX::COMFORT	Cold iron shackle, ball & chain	`Thu Feb 18 1993 14:54`	10
	This problem has been discussed before. I have circumvented the issue by checking an impossible situation, ie. the ethernet interface being down (if it is down, we can't talk to it) and generating an exception. My situation occurred running MCC on the VMS platform. I felt that the problem was caused by pings not refreshing the ARP cache and eventually the cache times out. Dave
4563.2	But on ULTRIX?	ZUR01::SCHNEIDERR		`Fri Feb 19 1993 03:33`	10
	Yes, But on ULTRIX it must be a leak or a bug somewhere. MSU does it o.k., but on MSU you have the possibility to set a reply counter, so you start a notification onli after i.e 2 or 3 pings. With MCC V1.2 you can't do that. So the PING gets lost somewhere. ON VMS maybe in your ARP cach and on ULTRIX ?????? Roland
4563.3	it works ok with the IP Poller of V1.3	TLSE01::SELLES	Pierre-Jean - Toulouse -France	`Fri Feb 19 1993 12:05`	12
	see note 4474 ; it is not solvable with v1.2 unless using occurs_n_times rules and wasting rules entries and cpu do do it . i tested the Ip poller of V1.3 and it works perfectly with no rule entry to waste and a background process doing all the work ( this is an ultrix version) cdmlt PJ
4563.4		ZUR01::SCHNEIDERR		`Thu Feb 25 1993 02:21`	7
	PJ, I thought to wait until V1.3 with the IP poller function. I told that to the customer. He is agree and we will wait. So lets hope we will get the V1.3 version very soon. Thanks Roland