ocf:pacemaker:ping

The ocf:pacemaker:ping resource agent is an OCF-compliant (Open Cluster Framework) agent designed to monitor network connectivity by using the standard ping utility. It is typically deployed in Pacemaker-managed clusters to verify that a target host (often another cluster node or a critical network endpoint) is reachable. If the ping check fails, the resource agent signals a failure, which can then trigger recovery or fencing operations.

Purpose

• Connectivity Monitoring: Periodically sends ICMP echo requests (ping) to a specified host to ensure that the network connection is active.

• Cluster Health: Helps in determining if a remote host or critical network path is available. In a cluster, this can be used to monitor node availability or to trigger automated responses if connectivity is lost.

• Failure Detection: Supports cluster failover by detecting when a host becomes unreachable, allowing the cluster to take corrective actions (such as resource relocation or node fencing).

Key Parameters

• host (required): The IP address or hostname of the target to be pinged.

• count (optional): The number of ping packets to send per check (default is typically 1 or 3, depending on the implementation).

• timeout (optional): The maximum time (in seconds) to wait for a ping response before declaring a failure.

• ping_options (optional): Additional options to be passed to the underlying ping command for fine-tuning its behavior.

Standard Operations

Like other OCF resource agents, ocf:pacemaker:ping supports a standard set of operations:

• start/stop: Although ping is primarily used for monitoring, the start and stop operations can be used to initialize or disable the monitoring process if required.

• monitor: The monitor operation periodically executes the ping command to verify connectivity. A successful ping returns an OCF_SUCCESS status, while a failure returns an error status.

• meta-data: Provides an XML description of the resource agent, including supported parameters and default values. You can view this metadata with a command such as:

  crm ra metadata ocf:pacemaker:ping

Example Usage in a Cluster

In a Pacemaker-managed cluster, you might create a ping resource with the following command:

pcs resource create ping_resource ocf:pacemaker:ping host=192.168.1.100 count=3 timeout=5 op monitor interval=30s

This command does the following:

• Creates a resource named ping_resource using the ocf:pacemaker:ping agent.

• Configures it to ping the host at 192.168.1.100 using 3 packets with a 5-second timeout.

• Sets up a monitoring operation that checks connectivity every 30 seconds.

Best Practices

• Parameter Tuning: Adjust the count and timeout parameters to suit your network conditions. A higher count can reduce false positives in unstable networks, but may also delay failure detection.

• Integration with Cluster Policies: Use the ping resource as part of a broader strategy to monitor node connectivity or critical network paths. Combine it with constraints and recovery policies in Pacemaker to automate failover.

• Resource Agent Metadata: Always review the metadata provided by the resource agent (via the meta-data operation) to understand the available parameters and defaults before deploying in production.

Conclusion

The ocf:pacemaker:ping resource agent is a lightweight and effective tool for monitoring network connectivity within a high-availability cluster. By leveraging standard ICMP ping operations, it provides real-time insights into the reachability of critical hosts, thereby helping to maintain overall cluster health and triggering automated recovery actions when connectivity issues are detected.