Acceldata
 Pulse
Search documentation… ⌘K

Last updated: Oct 30, 2026 15:11 UTC

Configure YARN Resource Fallback Values

Note

Configure YARN resource fallback values in advance to prevent incorrect resource allocation during an ad-events (NATS) outage.

Problem

When the ad-events (NATS) service is unavailable for an extended period, the PulseYARNMetrics Agent restores each NodeManager's original memory and vCore values from yarn-site.xml.

Previously, if the file could not be read, the agent assigned 8192 MB of memory and 4 vCores to every node, regardless of its original configuration.

Resolution

Configure fallback memory and vCore values in override.yml. The agent uses these values if it cannot read yarn-site.xml.

To configure the fallback values:

  1. On the Pulse node, go to the override.yml file:
/data01/acceldata/work/<cluster_name>/override.yml
  1. Under base, add the following parameters and specify the required values:
base:
  yarn_opt_default_memory_mb: "<memory_in_mb>"
  yarn_opt_default_vcores: "<vcore_count>"
  1. Run the following command to apply the changes:
accelo reconfig cluster

Note

The agent uses the memory and vCore values from yarn-site.xml when both are greater than zero. If the file cannot be read, it uses the fallback values configured in override.yml.