Writer instructions
Page title	For most spaces, this page must be titled Space announcements. For spaces with localized content, this page must be titled Space announcements l10n.
Purpose	Provide an announcement banner on every page of your space.
Location	Move this page outside of your home branch.
Guidelines	Adding-a-banner-announcement-to-a-space

Failover and cluster monitoring scenarios

The following failover and cluster monitoring scenarios are discussed:

- Failover and cluster monitoring by using the PATROL Agent
- Failover and cluster monitoring by using PATROL for Oracle

Note

If you want to use PATROL for Oracle to monitor RAC-managed instances, PATROL for Oracle must be installed on at least one node that contains a configured instance that is to be monitored. Failover monitoring cannot be used when monitoring a RAC-managed instance. To monitor the instance status within the RAC configuration, see Viewing and modifying HTML reporting

Failover and cluster monitoring by using the PATROL Agent

The PATROL Agent provides failover tolerance by operating in a cluster-specific mode (also known as the PATROL virtual agent). To install the PATROL Agent into a cluster configuration, see the BMC PATROL Agent.

Failover and cluster monitoring by using PATROL for Oracle

You can use PATROL for Oracle to monitor the following types of failover environments:

Active/passive environments
Active/active environment

Active/passive environments

You can use PATROL for Oracle to monitor an active/passive failover environment. To set up PATROL for Oracle and the PATROL Agent by using the PATROL Agent without the cluster-specific variables defined (non-clustered mode), use the following steps. The following figure illustrates this configuration. In Figure 6, Node 3 is the failover computer for Instance 1 on Node 1.

Install and start the PATROL Agent on both Node 1 and Node 3.
Install and configure PATROL for Oracle on each instance on Node 1 and Node 3.
On Instance 1 on Node 1, configure the failover system by using the*Failover Monitoring > Configure* command from the ORACLE_INSTANCE menu.

In normal operation, the failover instance (Instance 1 on Node 3) is in a standby mode, and the InstanceStatus, OracleStatus, and ConnectDB parameters are in warning mode on this instance. The primary instance (Instance 1 on Node 1) is running, and the InstanceStatus, OracleStatus, and ConnectDB parameters are in an OK state.

When a failover occurs, the InstanceStatus, OracleStatus, and ConnectDB parameters on Instance 1 on Node 1 go into warning or alarm (24x7). As Instance 1 on Node 3 starts and retrieves the redo log from the shared Oracle device, PATROL starts monitoring Instance 1 on Node 3. The InstanceStatus, OracleStatus, and ConnectDB parameters on Instance 1 on Node 3 go into an OK state.

In this operation, the PATROL history for each instance is interrupted. If you do not want to view the warning parameters on Instance 1, you can use the Monitor Enable/Disable menu command to stop monitoring the warning instance. To automatically disable or enable an instance so that it will not alarm on the secondary computer, use ORACLE_FAILOVER.km.

For information about configuring the failover monitor, see Failover monitoring.

Configuration with PATROL Agents not in cluster-specific mode

Active/active environment

PATROL for Oracle will also monitor the instance in an active/active environment. In this environment, which is also called a "workload balancing" configuration, all nodes have different instances. When a failover occurs, the remaining active node takes on the monitoring of all cluster applications. The following figure shows this situation. Node 1 contains the primary Instance 1, and is the failover computer for Instance 2. Node 3 contains the primary Instance 2, and is the failover computer for Instance 1. If Node 1 fails, Instance 1 fails over to Node 3, and the PATROL Agent on Node 3 starts monitoring Instance 1.

Note

If you are using the PATROL Agent in a non-virtual mode, you will not have continuous history data in a failover situation. The following figure displays the use of the PATROL Agent without the cluster-specific variables defined (non-clustered mode).

To set up PATROL for Oracle in this environment

Install the PATROL Agent on each node.
Install and configure PATROL for Oracle on each instance.
For each instance, configure the failover system by using the Failover Monitoing > Configure command from the ORACLE_INSTANCE menu.

In normal operation, Instance 1 on Node 1 and Instance 2 on Node 3 displays the InstanceStatus and OracleStatus parameters in an OK state. Instance 1 on Node 3 and Instance 2 on Node 1 displays the InstanceStatus, OracleStatus, and ConnectDB parameters in a warning state.

If a failover occurs on Node 1, Instance 1 on Node 1 goes down, and Instance 1 on Node 3 comes up. Instance 1 on Node 1 displays the InstanceStatus, OracleStatus, and ConnectDB parameters in a warning state. Instance 1 on Node 3 displays the InstanceStatus, OracleStatus, and ConnectDB parameters in an OK state.

In this operation, the PATROL history for the instance is interrupted. If you do not want to view the warning parameters on Instance 1, you can use the Monitor Enable/Disable menu command to stop monitoring the warning instance. To automatically disable or enable an instance so that it will not alarm on the secondary computer, use ORACLE_FAILOVER.km.

For information about configuring the failover monitor, see Failover monitoring.

Distributed environment configuration - local PATROL Agents only