What do I need to know to set up a Database Agent with High Availability?
In this article
Starting principles
The purpose of running agents in High Availability (HA) is to minimize disruption. When one agent fails or is experiencing downtime, then the other agent can continue to monitor and report.
At any given time, one Agent will be active while the others are in passive states. If one Agent goes down, the other Agents will, in turn, pick up the Database Collectors and begin reporting.
Agent location and naming for HA
To set up a Database Agent with High Availability (HA), we recommend installing each Agent on a different host. If the Agents must be on the same host, then they need to be installed into different folders.
The primary and secondary Database Agent should both have exactly the same name. For example, -Ddbagent.name=<same-name>.
|
DB AGENT
|
LOCATION
|
HOST
|
|
Primary
|
(host 1)
|
$ java -Ddbagent.name=abc -jar db-agent.jar
|
|
Secondary
|
(host 2)
|
$ java -Ddbagent.name=abc -jar db-agent.jar
|
Start the agents using the steps in Start and Stop the Documentation.
Crucial considerations for configuration
Keep in mind that all Database Agents in the HA configuration should:
- Be running at all times —
If the Agents are not running, then the Agents will not be visible on the Controller
- Have access to the monitoring databases which are defined in the Collector configuration
In addition, be sure that Collectors are configured to use the active Agent.
Additional Resources
