Dear,
I develop driver with SNMP trap receiver in which using possibility that alarms is generated when a trap is received.
Also, I use linking on bindings to allow clearing alarm with another trap. TrapOID definition is below. Sometime (very rare) have problem that trap for clearing alarm not clear active alarm (alarm stay active in window). In alarm history everything seems OK – Dataminer received trap for Critical and for Normal event. Received OID values for linking is OK. Did I missed something, or interpreted something wrong from Dataminer Development Library? What can be reason for alarm to be not cleared?
Thank you.
<TrapOID mapAlarm="TRUE|Severity:4:Critical,problem;Normal,recovery|Value:'[19]' [18]' '[14]' '[11]' '[9]' '[8]' '[6]' '[5]' '[3]'|Link:19,18,14,11,9,8,6,5" type="wildcard">1.3.6.1.4.1.56122.2.1.1.*</TrapOID>
This is a feature we rarely use anymore. Typically we create an alarm table which is maintained by the incoming traps. The oids can then fill the related columns.
With an alarm template the user can then chose if alarming is needed or not on the columns of choosing, which is a big advantage over the older trap feature.
That being said, there are some risks that need to be mitigated.
If there is a possibility for a lot of traps, we also typically make this a table which is not saved in the ElementData database table (see option ‘volatile’)
The danger with lots of traps is that it can result in a additional of load on the database.
Depending on the situation and the device, the other danger is that the table will grow without restrictions. It’s always necessary to prevent that scenario by removing entries automatically (either when they get too old, or with a limit on the amount)
I know it’s not the answer you were looking for, but the preferred method is with a table.
Hi Jeroen,
Thank you for advice. I will try to implement it with table.
Best regards,