[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

3358.0. "V1.2 Ultrix & Targetting on Iconic Map - BUG" by MUTTON::LAMB (Peter Lamb - GSG Santa Clara) Tue Jul 14 1992 22:24

Hello,

I have currently have a working DECnet sink setup between my Ultrix system
and my DECnet router.  Everything now seems to be working fine except 
targeting...

I am notifying on adjacency up/adjacency down messages and attampting to
use the new SBB 1.2 notification syntax which allows one to do this with
a single notification...  My notification request is as follows...

notify domain xxx entity list=(node4 router circuit qna-0 adjacent node 30.144),
events=(adjacency up, adjacency down) !severity=warning

To then cause these events to trigger the node with the correct severity
I then have a target assigned as follows...

assign target domain xxx event source=(node4 * circuit * adjacent node *),
event name ="adjacency down", managed object="node4 * circuit * adjacent node #1",
target entity = "node4 #1", target severity = critical

My problem is that no matter what, the targeted entity event ends up being
displayed on the iconic map as a WARNING instead of a CRITICAL.  Interesting
enough the targeted node4 does however get the WARNING instead of the router.

Note:  I know I have everything set up correctly because in the mcc_fcl mode
at the exact time that I get the incorrect targeting on the iconic map I
get the correct message in the mcc_fcl.  A copy of what I am receiving 
in the mcc_fcl is included below.

%%%%%%%%%%%%%% Event, 1991-07-15-01:15:16. %%%%%%%%%%%%%% [1]
Domain: LOCAL_NS:.physical.us.wea.wro1-1              Severity: Critical
Notification Entity: Node4 LOCAL_NS:.dna_node.mccwst
Event Source: Node4 LOCAL_NS:.dna_node.pjvax Circuit QNA-0 Adjacent Node 30.144
Successfully received events:
Event: Adjacency Down
 Adjacency Down
                                 Reason = Adjacency listener receive timeout
                  Adjacent Node Address = 30.144
                     Adjacent Node Name = MCCWST

                                                          

HELP!!!!  I have to get this working soon as I will be taking it to Europe
for demonstrations at a customer location!!

Thanks!

Peter 
T.RTitleUserPersonal
Name
DateLines
3358.1try NOTIFY without severityCTHQ4::WOODCOCKWed Jul 15 1992 09:4227
Hi Peter,

Is the NOTIFY request below is the severity really set to WARNING. I ask
because of the "!". In any case for the NOTIFY command, set NO SEVERITY. That 
is, let it default to INDETERMINATE.

The TARGETTING commands will then override the indeterminate on the TARGET
node. FWIW, you will probably need two target commands (one for each event)
to set one to CRITICAL and the other to CLEAR.

This works for us in a similar fashion for circuit events except the target
doesn't change. Give it a spin it might do the trick.

best regards,
brad...

>notify domain xxx entity list=(node4 router circuit qna-0 adjacent node 30.144),
>events=(adjacency up, adjacency down) !severity=warning

>To then cause these events to trigger the node with the correct severity
>I then have a target assigned as follows...

>assign target domain xxx event source=(node4 * circuit * adjacent node *),
>event name ="adjacency down", managed object="node4 * circuit * adjacent node #1",
>target entity = "node4 #1", target severity = critical


3358.2Tried it...PJWL::LAMBPeter Lamb - GSG Santa Clara MAIL=MUTTON::LAMBWed Jul 15 1992 12:2321
Hi,

Thanks for getting back to me so quickly!!!

The ! for severity level is what the iconic_map assign target/show
target displays.  Just for grins I will try creating the target at
the mcc_fcl level and see if it makes any difference but again it 
looks like the mcc_fcl has it right anyway...

I origionally had the notify set to INDETERMINATE. and only changed
it to warning to see if that new level also got xfered to the target.

In both cases either indeterminate or warning etc. the target receives
the level set by the notification not by the target assignment.

Thanks for the suggestions!!  I will play with it some more today...
If you have any more ideas please let me know!!!

Thanks!

Peter
3358.3Fixed it...MUTTON::LAMBPeter Lamb - GSG Santa ClaraWed Jul 15 1992 15:2719
Brad,

Setting no severity did the trick!!  It turns out that previously I had it set
to indeterminate but had the togle turned on such that it included a severity.

So I guess my next question is shouldn't targeting override the notification
severity level????

Additionally, I just noticed something else strange with notification...

I left the mcc_fcl mode running overnight with the notification

notify domain xxx event=(any config event)

and received the following errors sometime durring the night...  

DECnet IV am internal error - unfortunatly I lost the origional messages

Peter
3358.4haven't seen error with this kitCTHQ3::WOODCOCKWed Jul 15 1992 15:4720
Peter,

>So I guess my next question is shouldn't targeting override the notification
>severity level????

That would make sense to me but maybe the engineers had their reasons for
doing it this way. Right now I'm more worried about things that don't work
with the SSB rather than things that don't look right.

>DECnet IV am internal error - unfortunatly I lost the origional messages

I haven't seen the above error to date with this kit but I'm getting ALARM
errors which may be caused by DECnet IV am when ALARMS is doing global
wildcard rules. I'm actually getting so many errors I have rules turned off
so I'm only about 80% production. I'm thinking of punting back to T1.2.7 if
I can't get it fixed :-(.

best regards,
brad

3358.5targetting severity override strategyTOOK::CALLANDERMCC = My Constant CompanionTue Aug 11 1992 13:3427
    well engineers always have reasons, the question is simply "are the
    reasons any good?" Well we thought so but you decide (BTW we have
    updated the notification manual for the EMS release so as to
    better describe the targetting stuff).
    
    This is how targetting is decided:
    
    1) the entity returns the severity
    
    2) the user does an assign target to override the severity set
       by the entity
    
    3) the user tells the notify command what severity to associate with
       a notification, overriding any other value previously associated 
       with it.
    
    In other words, we took a top down approach for the overrides. The
    higher up in the food chain the more powerful your override.  So,
    if you say NOTIFY from the iconic map with an override severity of
    warning, EVERY event that is returned for that NOTIFY will have
    a severity of warning.  If instead you tell notify to pass up
    the severity (don't override with the PM switch) then you get the
    retargetting from the assign target, and if no assignments exist
    you get the entity supplied severity.
    
    hope this helps