Help with SNMP traps?

[scottwilkerson]

Are you sure these devices are setup to sent their traps to the Nagios server? Do they have a "test trap" function you could use to simulate a trap from them?

[jbennett]

As with the NRDS thread, I'm going to assume that if I already have these hosts set-up on my nagios box with basic SNMP checks, then that could be causing some of the problems?

ok - I am seeing 10 items show up in my unconfigured objects on my test box. All of these IPs that are showing up are already configured on my Nagios box, but are configured with a location identifiable host name instead of the IP as the host name. When I select all and tell Nagios to process them, they get added to my host list.

Then when I quick search based upon the IP, it pulls up BOTH hosts (same address). The original host shows that the SNMP Traps check has been completed and returns an OK state. The newly added host via unconfigured objects just sits as pending.

[scottwilkerson]

The newly added host via unconfigured objects just sits as pending.

It will until it receives another trap from the device

[jbennett]

But why am I not getting any other unconfigured devices? Only 10 out of nearly 400 devices that should be generating traps showed up under the unconfigured objects.

[scottwilkerson]

But why am I not getting any other unconfigured devices? Only 10 out of nearly 400 devices that should be generating traps showed up under the unconfigured objects.

Items will only show up in unconfigured objects if they send a trap to the XI server, and the XI server doesn't have the host/service configured. they would not show up until XI has received a trap from them.

[jbennett]

But why am I not getting any other unconfigured devices? Only 10 out of nearly 400 devices that should be generating traps showed up under the unconfigured objects.

Items will only show up in unconfigured objects if they send a trap to the XI server, and the XI server doesn't have the host/service configured. they would not show up until XI has received a trap from them.

I'm still lost.

If that's the case, then of the 390 other devices that I already have configured, why do those traps NEVER change state? We have a power outage at a location, the UPS switches to back-up battery and the secondary power source on the PDU goes dead. Both of those events should generate traps per the MIBs for each device, yet when I run a state history report, they never show up as a trap.

[scottwilkerson]

Can you verify the device sent the trap (on the device)?

[jbennett]

I have been looking into this and have found a number of settings burried deep on the PDUs that appears to have helped. I've gone through and changed these settings and a handfull of new items showed up in the Unconfigured objects settings.

This seems a bit backwards. If a piece of equipment alerts, but this is the first time it's alerted, I would have to go to unconfigured objects first to configure it. Then, and only then, would we be alerted to the issue.

If this were to happen while I'm out of the office, it would just sit there waiting to be configured.

I'm not sure I understand the purpose there? I will never have all of our items configured until every last one throws an alert in this case. Am I understanding that correctly?

[abrist]

If many of the devices are the same model, once one alarms and is recorded in unconfigured objects and you have added the object into the ccm, you could copy the config and change the ip of the new copied device to one of the yet-to-alarm devices. You either have to wait for the trap to be sent, or configure each device by hand. A number of devices have a management interface that you can use to force a trap to be sent, but I do not know if your devices have that feature.

[jbennett]

If many of the devices are the same model, once one alarms and is recorded in unconfigured objects and you have added the object into the ccm, you could copy the config and change the ip of the new copied device to one of the yet-to-alarm devices.

This helps a great deal. Thanks.