Volcano Batch Workload Management
Unified capability for managing Volcano batch workloads on Kubernetes. Combines Job, Queue, and PodGroup APIs to orchestrate distributed ML training, HPC simulation, big data processing, and scientific computing workflows. Designed for ML platform engineers, data engineers, and HPC cluster operators.
What You Can Do
MCP Tools
list-batch-jobs
List Volcano batch jobs in a Kubernetes namespace (ML training, Spark, HPC).
submit-batch-job
Submit a new distributed batch workload to Volcano (PyTorch, TensorFlow, Spark, MPI).
inspect-batch-job
Get detailed status and phase of a Volcano batch job.
cancel-batch-job
Cancel and delete a Volcano batch job and its associated pods.
list-cluster-jobs
List all Volcano batch jobs across all namespaces in the cluster.
list-scheduling-queues
List all Volcano scheduling queues and their resource allocations.
inspect-scheduling-queue
Get configuration, capacity, and status of a Volcano scheduling queue.
create-scheduling-queue
Create a new Volcano queue for multi-tenant resource partitioning.
list-podgroups
List PodGroups used for gang-scheduling coordination in a namespace.
inspect-podgroup
Get the status of a PodGroup to check gang-scheduling progress.