Description
There are several possible configurations of the event agent, controlled by the environmental variables:
CA_OPERA_NODE CA_OPR_PROXY CA_OPR_USEDB
Solution
When you install the Event agent, you have 2 options:
With the first option, you will install the 'lightweight' or 'dumb' agent, which forwards ALL the messages passing through it to an Event Manager.
Check "Get policies from a local DSB" then all messages from that agent will be sent to the manager, even after an 'opreload'. The message traffic will be high.
In that configuration, the following variables will be set:
CA_OPERA_NODE - <remote manager> CA_OPR_PROXY - unset CA_OPR_USEDB - <N>
With the second option, you will install the 'full' or 'smart' agent, that is able to process messages, execute actions on them and selectively forwards them to one or several event managers.
Uncheck "Get policies from a local DSB" then the local 'caoprdmn' will process local messages and selectively send some to the manager. Message traffic will be low.
In that configuration, the following variables will be set:
CA_OPERA_NODE - <local server> CA_OPR_PROXY - <remote manager>
When performing 'opreload', the agent will extract records from the database on the manager and build its own DSB.
Depending on the value of CA_OPR_USEDB, 'opreload' will or will not be called when starting the agent.
So, if CA_OPR_USEDB - string length 1, value <Y> the local DSB will be rebuilt from the remote database when the agent is (re)started.
If CA_OPR_USEDB - string length 1, value <N> the local existing DSB will just be read and loaded when the agent is (re)started.
Reloading the local DSB from either the remote DSB or the database is quite fast, generates little traffic, unless you do it very frequently, which is meaningless in production.
The 'smart agent' configuration will generate less network traffic than the 'dumb agent' configuration.