| 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 |
Hi,
VMS V5.4-3
DECmcc T1.2.7
I have been experimenting with wildcards in expressions using the
Node4 class. I am getting the following error even though the
attribute is known to the DNA4 AM.
Here's the rule:
!=======================================================================
! NODE_UNREACHABLE_DEV.COM
!=======================================================================
!
! =: Node unreachable :=
!
! Domain: PW.EH.DOMAIN.DEV
! simple: DEV
! Entity Class NODE4
! Instance *
!
DISABLE DOMAIN PW.EH.DOMAIN.DEV RULE NODE_UNREACHABLE_DEV
!
DELETE DOMAIN PW.EH.DOMAIN.DEV RULE NODE_UNREACHABLE_DEV
!
CREATE DOMAIN PW.EH.DOMAIN.DEV RULE NODE_UNREACHABLE_DEV -
EXPRESSION = (NODE4 (some router) REMOTE NODE * -
STATE = UNREACHABLE , AT EVERY 00:01:00) ,-
ALARM FIRED PROCEDURE = MCC_COMMON:MCC_ALARMS_MAIL_ALARM.COM ,-
ALARM EXCEPTION PROCEDURE = MCC_COMMON:MCC_ALARMS_MAIL_EXCEPTION.COM ,-
ALARM FIRED PARAMETER = "ECON02::LANMON" ,-
CATEGORY = "Node Is Unreachable from Router EI2RTR" ,-
DESCRIPTION = "PWA_MCC_FILES:NODE_UNREACHABLE.TEXT" ,-
BATCH QUEUE = "ALARMS$BATCH" ,-
SEVERITY = CRITICAL
!
DEASSIGN TARGET DOMAIN PW.EH.DOMAIN.DEV -
event source = DOMAIN PW.EH.DOMAIN.DEV RULE NODE_UNREACHABLE_DEV
!
ASSIGN TARGET DOMAIN PW.EH.DOMAIN.DEV -
event source = DOMAIN PW.EH.DOMAIN.DEV RULE NODE_UNREACHABLE_DEV,-
event name = "Any Notification Event",-
managed object = "Node4 (some router) remote node #1",-
target entity = "NODE4 #1",-
target severity = CRITICAL
And I get the following error:
Subj: Notification of exception " MCC 0 ALARMS RULE
NODE_UNREACHABLE_DEV 30-APR-1992 19:19:35.31"
Rule name: MCC 0 ALARMS RULE NODE_UNREACHABLE_DEV
Domain: Domain PW_NS:.PW.EH.DOMAIN.DEV
Occurred at: 30-APR-1992 19:19:35.31
Category: Node Is Unreachable from Router (some router)
Description: PWA_MCC_FILES:NODE_UNREACHABLE.TEXT
Severity: Indeterminate
Expression: (NODE4 (some router) REMOTE NODE *
STATE = UNREACHABLE , AT EVERY 00:01:00)
Exception: Rule could not be evaluated as the attribute was not
returned by the Entity.
Any ideas ??
Thanks.
Joe
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 2907.1 | 1 minute too fast | ICS::WOODCOCK | Fri May 01 1992 08:47 | 30 | |
> Expression: (NODE4 (some router) REMOTE NODE *
> STATE = UNREACHABLE , AT EVERY 00:01:00)
> Exception: Rule could not be evaluated as the attribute was not
> returned by the Entity.
>
> Any ideas ??
Hi Joe,
Just a guess. The router database for remote nodes is probably very large, most
likely around 1000 nodes at least. Your polling rate is for every minute. There
is no way MCC will be able to process all the remote nodes in a minute. You
could issue the command with fcl:
show node4 router remote node * state
See how long this takes, grab a coffee (maybe even lunch).
The second problem is that when the above SHOW command is issued the router
only sends back the status of REACHABLE nodes with a wildcard. This is true
for both MCC and NCP. Only when you specify non-wildcard node number (that you
know is down, or the address is skipped with the above command) will it return
UNREACHABLE. Therefore, MCC would have to collect all the address which fit
the wildcard, then issue a seperate command for each address to actually find
the UNREACHABLES. I don't know if MCC does this within their alarms.
best regards,
brad...
| |||||
| 2907.2 | rule isn't practical | ICS::WOODCOCK | Fri May 01 1992 09:58 | 26 | |
> Expression: (NODE4 (some router) REMOTE NODE *
> STATE = UNREACHABLE , AT EVERY 00:01:00)
> Exception: Rule could not be evaluated as the attribute was not
> returned by the Entity.
>
> Any ideas ??
Hi Joe,
You'd think I could get all my thoughts together into one note. Anyway, the
above rule will not work with REMOTE NODE * even if you were to fix and/or
figure out what is happening based on .1 comments. Actually I should say that
it would not be practical, I don't think. For example if you had 23 active
nodes in your DECnet area, and you got the above rule to work the following
would occur. Assuming even that all 23 nodes were up, MCC is still going to
try to resolve all the rest of the addresses for that area. As far as the
router is concerned the other 1000 nodes within the area are UNREACHABLE.
Hence, MCC would fire for every non-active node in the area. In this case
the rule would fire 1000 times every polling period. The problem is that the
ROUTER doesn't know what addresses are active (but down), and which addresses
are simply not used at all. MCC is getting the info from the router, G-I-G-O.
best regards,
brad...
| |||||
| 2907.3 | Thanks. I'll try GETEVENT | MERIDN::ATTERBERRY | Fri May 01 1992 10:36 | 10 | |
Thanks Brad,
I did not realize how MCC would handle such a request. Thanks for the
info. I think I might try using the GETEVENT functionality. The only
thing I don't like about that is that I will have to install the DECmcc
Director license and BMS software on one of the routers. It may be
worth it to get around polling. I really want to keep the DECmcc
system an end node.
Joe
| |||||
| 2907.4 | try another router | ICS::WOODCOCK | Fri May 01 1992 13:15 | 15 | |
Hi Joe,
> I did not realize how MCC would handle such a request. Thanks for the
> info. I think I might try using the GETEVENT functionality. The only
> thing I don't like about that is that I will have to install the DECmcc
> Director license and BMS software on one of the routers. It may be
> worth it to get around polling. I really want to keep the DECmcc
> system an end node.
If you have privs for ANY router you can set up a sink for that router to
send the events to your MCC end node rather than changing your node to a
router.
good luck,
brad...
| |||||