Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations strongm on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Generate alarms from 3com switch

Status
Not open for further replies.

wenger2001

Instructor
Jan 2, 2007
3
GB
Hi,

I have a heavily switched 3com network, using 3com 4200's. I can manage switches using nnm, however alarms aren't being generated when a switch fails. Reading through the documentation there doesn't seem to anyway to generate snmp traps for an interface failure on a 3com switch.

Surely it must be possible for nnm to identify a switch failure?

Thanks

Dan
 
Hi wenger,
Just a quick question, do you really need to know when any interface/port in the switch goes down? Or can you have one single interface with an IP address for management and when this one fails you can have a Node Down alarm?

If the switch fails, or anything in the switch fails, it should send its own snmp traps to NNM if snmp is properly configured in the switch.

Please clarify what you mean by switch failure and we can provide some additional help!

Hope this helps a little wenger.
Bests,
Carlos.

vlan52
The end of wisdom is freedom. The end of culture is perfection. The end of
education is character. The end of knowledge is love.
 
vlan52,

Thanks for the reply. One single interface can go down with a management ip address to generate a node down alarm.
It doesn't really matter which interface.

But I can't seem to get this to work. Help much appreciated.

Thanks

wenger
 
Excellent,
Now, you should have all the interfaces set us unmanaged except for the management interface/IP address.

The node should have that IP as the only reference for its name under NNM. When you discover the node it should discover it using that IP.

NNM already has an "internal trap" as node_down, this is a generic trap, so as long as the switch has only one interface (the management interface) as managed, when the management interface goes down, you'll receive a Node Down message, internally generated by NNM.

You can also have un Node Up message when the interface come back up again.

There is no need for the switch to send SNMP notifications under this scenario. Still, if you want to use the rest of the SNMP information from the switch, the 3com should have the definition of the snmp server (the Ip of the NNM), the same read-only community as declared under NNM.
NNM should have loaded the particular set of MIB for 3com switches of course.

Please let us know if there is anything else you need Wenger!

Bests.

vlan52
The end of wisdom is freedom. The end of culture is perfection. The end of
education is character. The end of knowledge is love.
 
Hi,

Thats what I thought but it doesn't seem to be working. I have the following scenario

router A--switch B---switch C

If I disable the B-C link interface on switch A, I do not get a nnm ov_node down alarm, in fact I get nothing at all.
Even if I try right clicking on the object and go ping in nnm it fails but doesn't bring up a node down alarm.

surely it should give an alert?

Thanks
 
Hi,
Thanks for your questions!
Just a few things...

When you say "disable" the B-C interface do you mean unmanaged under NNM or just to disable the interface on the router/switch as an admin would do?

If you unmanage it from NNM, it won't show any alarms, because it is supposed to be a useless one.

For NNM to know the right status of each component, router A should have one mgmt interface UP and managed, also the ports connecting to switch B, so it can verify the status of B and then B can have its own management interface + the port to switch C as managed under NNM.

When you right click on a node to ping it, it is going to resolve the name for the node, to its IP address and then run the icmp tests. Is it resolving the name? Or is it using just the IP instead?

Remember NNM has a threshold configured in order to make sure a node shows as down after a certain number of polls/pings, we must make sure that threshold is set to an acceptable value, let's say 2 minutes i.e.

All the managed nodes should have name resolution configured for the exact name you want under NNM or Node Bank and the management IP address you are using, this is VERY important.(can be done through DNS or local resolv file, I would recommend DNS dedicated for management purposes).

Also, in the Alarm Browser, you can check on Event Configuration, for the particular event (Node Down or IF_Down) to make sure these are configured to log AND show alarms, perhaps they are only being logged and nothing else. Also make sure it say "All sources" there, instead of particular set of pre configured nodes.

Let's see if this helps a little!





vlan52
The end of wisdom is freedom. The end of culture is perfection. The end of
education is character. The end of knowledge is love.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top