yarn is not utilizing the cluster resources and jobs are waiting in the accepted state.
Can you please share the following?
1. Resource Manager (RM) logs
3. Screenshot of the Dynamic Resource Pool Configuration
4. Screenshot of the Queue Representation in RM Web UI
Hope, the above-mentioned artefacts would be sufficient for initial analysis of the issue.
We need to see the cluster Resource usage during the time frame when the jobs are in accepted state. If the entire memory and allocated vcores has been used the job does not have suffiicient resources to run the application. We need to look for the RM webui and Scheduler screenshot of all the queues to view the usage of the cluster.