The purpose of this document is to describe the inner workings of the "IBM Director" trap handler and show how to use it.
More information about Trapper can be found in the Trapper administration manual.
- Basic understanding of SNMP-traps and the Trapper functionality of op5 Monitor
- Installed the "IBM Director" trap handler" and configured a matcher for ".126.96.36.199.188.8.131.52.159.201*"
- A host in op5 Monitor with the same name or alias as the origin system and with a passive service called "IBM-Director"
- A host in op5 Monitor called "IBM-Director" with a passive service called "IBM-Director"
Configuring SNMP-traps in IBM Director
Before we start to look at the trap handler, we need to configure IBM Director to send SNMP-traps:
- Open the IBM Director web console
- Open "Automation" in the left side menu
- Select "Event Actions", then click on the "Create" button
- Enter the name and IP-address of your op5 Monitor host
- Open "Event Automation Plans" in the left side menu
- Select "Send SNMP Trap" and click the "Edit" button
- Add your op5 Monitor host from "Event Actions" and save the configuration
Matching process overview
The trap handler looks at the "managed object name"-field in the trap to determine which system the alert originated from.
It uses this field to map it against a host in op5 Monitor with the same name and a service called "IBM-Director".
If it can't find this host, it will instead appear under the "Subservices" section in "SNMP traps".
The trap handler can also try to map the "managed object name"-field against host aliases, see more under the "Advanced usage" section below.
Successful mapping against the host "ibmsystem.example.com" and service "IBM-Director"
If the host and service mapping was unsuccessful, it can still be found under the "Subservices" section
Testing the trap handler
You can test the trap handler by sending your own traps with the command line utility "snmptrap", available on the op5 Monitor host.
The following command will send a SNMP-trap to 127.0.0.1, which should make the trap handler put the service in a "WARNING" state:
This command should put the service back in to a "OK" state:
The trap handler has a function to map the "managed object name"-field against a host alias and remove domains if needed.
This can be very useful if you don't have matching host names or haven't included the FQDN of the host in op5 Monitor.
If the mapping still isn't successful, it will put the trap on a generic "catch-all" host called "IBM-Director" - this is done to ensure that you always get notified when alerts are received.
However, this function requires you to use a script called "gen-mapping.lua" to create a key-value list of hosts and aliases from the database.
The script should be run every time you add new hosts and the module updated, so we suggest executing it hourly with cron:
You will also need to change the "useAliasMapping" variable in the handler to "true".
Technical description of the mapping process
The following is an example of the content of a trap received from IBM Director:
This section in the trap handler shows how the mapping of the OIDs is done:
The "map" function tries to match "trapManagedObjectName" with a host or alias, generated by the "gen-mapping" script.
The following code maps the trap to a service called "IBM-Director" on a host found by the "map" function if it is enabled.
It also sets the status message of the service to the contents of the "trapText" field:
The mapping of service severity status is using the "trapSeverity" field:
Extend the functionality
You can easily improve and extend the functionality of the trap handler. For example, it currently maps all traps to one service called "IBM-Director.
If you want to separate traps regarding the physical memory status or similar, you can use the "trapType" field and look for "MPA.Component.DIMM":