About MSDP Scaleout
MSDP Scaleout is based on MSDP. It empowers MSDP with high resilience and scalability capabilities to simplify management and reduce total cost of ownership.
It runs on multiple nodes to represent a single storage pool for NetBackup and other Cohesity products to use. You can seamlessly scale out and scale up a MSDP Scaleout on demand. MSDP Scaleout automatically does failure detection and repair in the background.
The core MSDP services run on each node to expose the storage optimized services, and manage a part of the cluster level data and metadata. Each MSDP Scaleout node is called MSDP engine.
Following are the MSDP Scaleout components:
MDS (MetaData service)
MDS is an independent and stackable service that provides a single system view of MSDP Scaleout. It's an etcd cluster running inside the MDS pods. These pods run on different AKS or EKS nodes. The pod name has a format of <cr-name>-uss-mds-<1,2...>.
The number of pods that get created depends on the number of MSDP Scaleout engines in a cluster. These pods are controlled by the MSDP operator.
1 or 2 MSDP Scaleout engines: 1 pod
3 or 4 MSDP Scaleout engines: 3 pods
5 or more MSDP Scaleout engines: 5 pods
MSDP Scaleout Controller
Controller is a singleton service and the entry point of MSDP Scaleout that monitors and repairs MSDP Engines. It controls and manages the application-level business of the MSDP Scaleout. The Deployment object name has a format of <cr-name>-uss-controller. It is controlled by the MSDP operator.
MSDP Scaleout Engine
MSDP Engines provide the ability to write deduplicated data to the storage. The name of a MSDP engine pod is the corresponding FQDN of the static IP that is specified in the CR. Each MSDP engine pod has MSDP services such as spad, spoold, and ocsd running. They are controlled by the MSDP operator.