[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference netcad::hub_mgnt

Title:	DEChub/HUBwatch/PROBEwatch CONFERENCE
Notice:	Firmware -2, Doc -3, Power -4, HW kits -5, firm load -6&7
Moderator:	NETCAD::COLELLADT

Created:	Wed Nov 13 1991
Last Modified:	Fri Jun 06 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	4455
Total number of notes:	16761

1079.0. "Starting HUBwatch TRAP processes ???" by ZTOIS1::VISTA (Renato VISTA, MCS/SPS France/Strasbourg) Tue Jun 07 1994 15:56

    

Hi,

Working with HUBWATCH V3.0 for VMS (5.4-3), I'm trying to received SNMP traps
sent by DEChub900, where SNMP Trap Destination Addresses have been previously
defined.

Following the documentation/user's guide, I've created 
HUBWATCH_ALARMED_AGENTS.DAT using Applications/Alarm Setup Option

#
HUB_ID     :     hub900
        IP                     : 16.188.48.120
        COMMUNITY              : public
        POLLING_INTERVAL       : 20
        NOTIFY_TRAPS           : TRUE
#
HUB_ID     :     da90
        IP                     : 16.188.48.123
        COMMUNITY              : public
        POLLING_INTERVAL       : 20
#
HUB_ID     :     ds900
        IP                     : 16.188.48.122
        COMMUNITY              : public
        POLLING_INTERVAL       : 20
        NOTIFY_TRAPS           : TRUE
#
HUB_ID     :    pb900
        IP                     : 16.188.48.124
        COMMUNITY              : public
        POLLING_INTERVAL       : 20
#


The LOGfile has also been created via VIEWER window/search Option 
(HUBWATCH$ALARMS_LOGFILE.DAT).

Then, the POLLER has been started... So, all the expected processes seems to 
be running...

Doing AuthenticationFailure using MCC SHO SNMP <hub900> ALL CHAR, BY PASS
<badCommunityName> command from another ULTRIX system, I've got the expected
messages on MCC/ULTRIX system, ie

%%%%%%%%%%%%%% Event, 1994-06-07-16:40:44. %%%%%%%%%%%%%% [1]
Domain: LOCAL_NS:.openlink_hub900                     Severity: Indeterminate
Notification Entity: SNMP LOCAL_NS:.snmp.hub900
Event Source: SNMP LOCAL_NS:.snmp.hub900
Event: authenticationFailure
An authenticationFailure trap was received:
                             enterprise = "1.3.6.1.4.1.36.2.15.10.3.1"
                             agent-addr = 16.188.48.120
                           generic-trap = authenticationFailure
                          specific-trap = 0
                             time-stamp = 265846


BUT, I've GOT no message in the HUBWATCH/VIEWER window.
STOPPING HUBwatch processes, and running MCC/VMS on the same system, I've
got the expected messages

%%%%%%%%%%%%%% Event,  7-JUN-1994 17:42:05 %%%%%%%%%%%%%% [2]
Domain: LOCAL_NS:.demo_openlink                       Severity: Indeterminate
Notification Entity: SNMP LOCAL_NS:.snmp.hub900
Event Source: SNMP LOCAL_NS:.snmp.hub900
Event: authenticationFailure
An authenticationFailure trap was received:
                             enterprise = "1.3.6.1.4.1.36.2.15.10.3.1"
                             agent-addr = 16.188.48.120
                           generic-trap = authenticationFailure
                          specific-trap = 0
                             time-stamp = 265845


IS THERE a KNOWN problem in the STARTING phase of HUBwatch 
POLLER/LOGGER/NMS_EVENTD processes ?? Or do I make any mistake in the starting
sequence/definitions ??

Thank you for your replies.
Renato


PS : I've of course checked that NMS_EVENTD is starting when MCC_TCPIP_SINK 
process is not running at the same time...

T.R	Title	User	Date	Lines
1079.1	some suggestions from alarms developer	NACAD2::HAROKOPUS	`Fri Jun 10 1994 15:22`	21
	Renato, There are two things I can tell you to try for now. Hopefully one or the other will work. In release 3.0 (will be fixed in 3.1) we didn't get the account privs right. Try to run this with all your privs turned on. It also behaves differently if you start the Viewer from within HUBwatch than if you start it using command line syntax. So you may want to try both and see if one or the other works for you. This too will be fixed in 3.1 The other thing you can try is to just bring up the Alarms Setup window again. It looks like you've got things set up correctly in the file. You just need to make sure that the poller is aware that your workstation is to receive traps. You appear to have verified this for you station using MCC but you might also want to use the db900 console to verify that the workstation ip address is in there. This should happen automatically for you as you bring up the Setup window but you can easily verify that it did by using the console.
1079.2		NACAD::WILSON	`Mon Jun 13 1994 16:06`	72
	Renato, I apologize for taking so long in answering your note. But here goes... I'm not sure that you've followed the steps correctly. In your note you said: >> The LOGfile has also been created via VIEWER window/search Option >> (HUBWATCH$ALARMS_LOGFILE.DAT). The logfile is not created by going into the viewer window. The logfile is created by the HWLOGGER process when it receives a trap that is to be logged. >>Then, the POLLER has been started... So, all the expected processes seems to >>be running... After the HWPOLLER process has been started (by either the menu pulldown: Application->Alarms->Start Poller; watch/poller start; or by pressing the Start Poller button in the Alarm Setup window) have you entered the Alarm Setup window? You must first start the Alarm Poller (process name: HWPOLLER) and then "register" with the poller process the hub(s)/module(s) you are interested in. By bringing up the Alarm Setup window you will register those agents which have Traps "turned on" (ie the pushbutton is green and says ON). The Alarm Poller is configured to receive all Traps that are sent to your workstation. But, in order for you to see the trap in the Alarm Viewer, the module/hub you are interested in must be registered with the Alarm Poller. If you don't "register" the modules, the module will send out the trap to any pre-configured addresses but the poller will not know that you are interested in being notified about any trap messages from that module. Below is an explanation of how to "register" a module/hub with the Alarm Poller and what happens when you "register" it. "Registering" a module/hub can occur several ways. One way is when you actually add an agent through the Add window of the Alarm Setup window. Another is when you turn Traps "On" and press the Ok or Apply button. Yet another is when you enter the Alarm Setup window. Any hub/module entries with Traps "On" will be "registered". Once you've "registered" a module/hub with the Alarm Poller a few things happen. 1) a message is sent to the Alarm Poller process - which says add this module/hub. Here is an IP address, community string, name, and the items we are interested in (ie Traps). 2) The Alarm Poller takes the community and IP address you've provided and uses this to set your workstations IP address in the trap table for the module/hub specified. The Alarm Poller uses C functions to determine what your workstations IP address is. 3) Then, if Traps is On, the module/hub is added to an internal list. When a trap is received by the Alarm Poller this list is searched. If a match is made between the trap IP address and an address in the list, then the Trap is forwarded to the Alarm Viewer and Alarm Logger. I hope this explanation clears up some of the confusion (and doesn't cause more). If you still have questions or problems feel free to send me mail. Karen Wilson HUBwatch developer