Configuring Ingres High Availability Option for Linux
Configuring the OCF Resource Agent with Pacemaker
To provide high availability in a Linux Cluster environment, Ingres provides an Open Cluster Framework (OCF) resource agent script template that treats an Ingres instance, together with the Management Server, as a resource.
You can run a script to make an instance-specific version based on the provided template, and then install it.
To configure the OCF Resource Agent
1. Locate the template, named template.ocf, is located in $II_SYSTEM/ingres/files/ocf.
2. Source the environment file $II_SYSTEM/ingres/.ingIIsh. (The following example assumes an operating system user ID of ingres and an instance ID of II.)
source ~ingres/.ingIIsh
3. Run mkocf as root. This creates a version for the specific Ingres instance in the ocf directory. For example, if the instance ID is II, the script would be named ingresII.
mkocf
4. Use mkocf to install the script
mkocf -i
The script is copied to the /usr/lib/ocf/resource.d/heartbeat directory and is made executable.
5. To install the script on the other nodes in the cluster, issue:
# mkocf -i -n node2
Note: Steps 3 and 4 can also be done manually.
The script is recognized as a possible resource under Pacemaker
# pcs resource list | grep ingres
ocf:heartbeat:ingresII - Manages Ingres DBMS instance II
6. Register the script as a managed resource.
# pcs resource create IngresII ocf:heartbeat:ingresII
# pcs enable IngresII
The script implements the mandatory actions of start, stop, meta-data, validate-all, and monitor. It also returns an appropriate code for known unimplemented actions such as migrate-to and migrate-from.
Activity is logged to $II_SYSTEM/ingres/files/ocf_ra.log. This log “rotates” to a timestamped version, which can be removed, archived, or compressed when the log reaches a maximum number of lines. The maximum number of lines is configured by the ocf_ra_max_log_size configuration parameter, which defaults to 100,000 lines.
Last modified date: 08/28/2024