T.R | Title | User | Personal Name | Date | Lines |
---|
3421.1 | Notifications on STATION, REFERENCE or DOMAIN's | COPCLU::SORENC | S�ren H Christiansen - (7)857-2107 | Fri Jul 24 1992 05:17 | 25 |
| Hi,
I have with some luck used two different schemes.
1. As the PC is hardly registered as NODE4 (NML is never there) is can
(most of the time) be registered as a STATION. From the REACHABILITY
CHANGE event from a DECnet router, you will be notified that the PC is
up and down. Then you can create an alarm rule towards the PC
(STATION). If you do not like this, try no 2.
2. If you have DECnet nodes without NML present, you may register these
entities as REFERENCE (I have also used DOMAIN's, with some reasonable
icons). Now after the REACHABILITY CHANGE event you go to the router to
find if the node is UNREACHABLE, if so you make an alarm rule on the
node (i.e. the REFERENCE or DOMAIN entity), which always evaluates
TRUE.
I have only done this for DECmcc V1.1, and used a modified
TARGET_ENTITY command file !
One of my cases for such a scheme involved non-DEC DECnet routers, i.e.
Proteon routers, without SNMP support on the management station.
S�ren
|
3421.2 | | FOUR62::LICAUSE | Al Licause (338-5661) | Fri Jul 24 1992 10:18 | 14 |
| RE:.1 thanks very much for the suggestions.
However, I tried the station registration and have not been able to get MCC
to see any of two DEC pc's and one MAC in our demo center. So either I'm doing
something incorrectly or something is broken.
Also, your second suggestion......how to you go about determining what action
to take when a load event or load failure occurs? Or are you periodically
polling with alarms or checking the log file?
I would assume your using a command procedure to parse the output of the
notification event to determine which PC rebooted?
Al
|
3421.3 | No constant polling | COPCLU::SORENC | S�ren H Christiansen - (7)857-2107 | Fri Jul 24 1992 12:26 | 16 |
| I have only looked into the issue if a DECnet node is reachable (=
available ?), and as discussed elsewhere is these notes, I am using the
event Node Reachability Change event 4.14. For NODE4's which are
represented only by a REFERENCE or DOMAIN entity, you poll the router
just ONCE after the 4.14 event to check the actual reachability (seen
from the router), and then make the alarm. (I do not know, how your
load events and load failure events correspond to the situation of the
node actually being reachable/available or not).
The TARGET_ENTITY command file gives an example on how to retrieve the
needed information (as I recall event 4.14 holds the DECnet address for
your target node).
Regards,
S�ren
|
3421.4 | Problem fixed? ....maybe | FOUR62::LICAUSE | Al Licause (338-5661) | Mon Jul 27 1992 12:12 | 18 |
| Soren,
Thanks very much for your input. It wasn't until I read the Ethernet AM Users
guide that I discovered that both the Ethernet Address and the Ethernet form
of the DECnet address are required to register a PC in DECmcc.
IT SHOULD BE NOTED THAT NEITHER FIELD IS INDICATED AS BEING REQUIRED ON THE
ICONIC MAP PM AS MARKED WITH AN ASTERISK.
I would think that at least ADDRESS is a required field for registering
any Ethernet station.
With regard to notification events for stations, it appears to be pretty limited
to all or nothing. I'll play with it somemore and see what happens.
It looks as though an alarm will be the only thing to satisfy most situations.
Al
|