[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

3057.0. "communication has been interrupted" by ZUR01::SCHNEIDERR () Fri May 22 1992 08:18

Hello,

We poll bridges 600. All output is written in logfiles. We see often the message
"Communication with target has been interrupted". We think that it could be 
a kind of timeout in the RBMS protocoll. 

a) Does somebody know more aboutr this message or has seen this, too?

b) Are there parameter we could change in the RBMS protocoll (timeout, retry 
   limit and so on)

c) What could be the reason for that behvior?


Thanks and regards Roland

T.R	Title	User	Personal Name	Date	Lines
3057.1		QUIVER::HAROKOPUS		`Fri May 22 1992 13:28`	14
	Hi Roland, >"Communication with target has been interrupted". This usually happens when there is another process trying to poll the same bridge with RBMS at the same time. Are there any other active processes on your system using RBMS to talk to bridges? (e.g. DECelms, other DECmcc processes etc.) Regards, Bob
3057.2		SLINK::CHILDS	Ed Childs	`Wed May 27 1992 10:40`	3
	That message is a timeout message. Are you doing lots of bridge management functions on your network? Do you have a lot of Alarms set up for bridges?
3057.3		ZUR01::SCHNEIDERR		`Tue Jun 09 1992 05:18`	34
	Hi ed and Bob, here some answers to your questions: >>> Are you doing lots of bridge management functions on your network? About 200 - 220 requests in 5 minutes, only with one MCC station >>> Do you have a lot of alarms set up for bridges? About 250 >>> Does your network have unusually high traffic? 3-5 % multicats 8-10% ethernet longtime contention 30-50% ethernet peak contention 300 frames/sec average >1000 frames/sec peak >>> Are there other active processes on your system using RBMS to talk to the >>> bridges? With backgroundprocess for alarms and sometimes wia MCC iconic map >>> Does an other process poll the same bridge with RBMS protocoll? no Regards Roland
3057.4		ZUR01::SCHNEIDERR		`Fri Jun 19 1992 05:22`	7
	Have you seen the requested infos? Is there anithing other what I can do, or do we just have to accept that the communication to the bridges interrupt. If we have to, how shall I tell that to a customer? "sorry, but we can't fix your problem!" Thanks and regards Roland
3057.5		SLINK::CHILDS	Strike a poseur.	`Fri Jun 19 1992 10:40`	10
	I've changed the Bridge AM code to use a longer timeout. This change will be in the DECmcc ELM V1.2 kit. Based on the info you gave, you are polling one bridge in your network about every second (250 bridges every 300 seconds). This is a lot of network management traffic. You are probably getting timeouts when you poll bridges at the far ends of your extended LAN. The new timeout will be 2 retries of 15 seconds. I think this should fix the problem you are seeing.
3057.6		SUBWAY::REILLY	Mike Reilly - New York Bank District	`Mon Jun 22 1992 15:19`	6
	Ed, Thanks for pushing this thru for the 1.2 kit. The timeout and 'communication has been interrupted' errors are currently generating 85% of the exception messages on my customers DECmcc system. - Mike
3057.7	will wait to 1.2	ZUR01::SCHNEIDERR		`Tue Jun 23 1992 10:25`	3
	We will test it on 1.2 Thanks and regards Roland