Bare Metal Orchestrator logs are automatically collected and made available in a centralized OpenSearch dashboard that is deployed on the Global Controller. A unique Bare Metal Orchestrator ID is used to associate logs with a specific Bare Metal Orchestrator cluster deployment.
When you install Bare Metal Orchestrator, a unique ID is assigned to the cluster. All logs and events that are associated with the Bare Metal Orchestrator cluster reference that ID.
If you are monitoring multiple Bare Metal Orchestrator clusters, you can track which logs and events belong to each cluster using their unique cluster ID. Although a remote possibility, it is highly unlikely for the same cluster ID to get assigned to more than one Bare Metal Orchestrator cluster. For high availability (HA), the cluster ID is assigned to the Global Controller node (CP1) and is maintained during fail over scenarios.
You can see the Bare Metal Orchestrator cluster ID on the About screen using the Bare Metal Orchestrator web UI or using the bmo version -b
command.
Log aggregation
Fluentd performs log aggregation and an OpenSearch dashboard displays these logs. When Bare Metal Orchestrator is deployed, an OpenSearch dashboard pod and a Fleuntd logging agent pod are automatically created on the Global Controller.
OpenSearch only exists on the Global Controller, and Fluentd is deployed at all sites to collect logs. When a worker site is created by the Global Controller, a Fluentd pod is deployed on the worker site. Fluentd pushes logs collected at the worker site to OpenSearch.
Fluentd collects the following Bare Metal Orchestrator application logs from all sites:
- file_name
- message
- log_level
- service_name
- site
- timestamp
You can filter and query these logs in OpenSearch. Use these logs to view errors in a configuration profile, errors from a site, or all errors during a specified time frame.
To use the Bare Metal Orchestrator CLI to export logs and audit logs for a specific date or date range, see Export logs.
Access the OpenSearch dashboard
Run the following command to retrieve the internal IP address of the Global Controller:
bmo get node | grep <global controller site name or nodeName>
For example:
bmo get node | grep gc
Where gc is the default site name of the Global Controller. The OpenSearch dashboard endpoint URL is http://<global_controller_internal_ip>:31118. Use this endpoint URL to access the OpenSearch dashboard in a web browser.