Different storage policies can be selected according to different business types. After a virtual storage volume is created, the virtual storage will automatically generate built-in storage policies based on commonly used storage configurations. If the built-in policy does not meet the requirements, you can add a new storage policy to customize storage-related configurations such as replicas, automated QoS, and stripe width.

The following are the steps to add a new storage policy.

Steps:

Log in to the SCP management platform. Go to Resource Center to select the resource pool, click Storage >Virtual Storage > Storage Policies click New 3, fill in the policy name and description and select the replicas, stripe width, automated QoS and Replica Defrag, you can also set the new policy as a default policy.

Replicas: Define the number of replicas for the storage strategy.

Stripe width: Striping is to divide data into small pieces and save them to different physical disks. The stripe width reflects the number of physical disks that can be used at the same time. There is a normal curve relationship between the business service performance and the stripe width, and an appropriate stripe width is required to maximize IO performance. Recommended select Adaptive in most cases. In an all-flash scenario, you can set the number of stripes to the number of SSDs on a single node, to achieve the concurrent performance of all SSDs on this node.

Automated QoS: Automated QoS restrains the capability of getting tiered storage space of cache disk. High-priority virtual disks are more likely to get cache disk space (applicable to datastore with cache disks only). Recommended set QoS to high level of performance for only a few business scenarios with high performance requirements and use medium level of performance for other business scenarios.

Example: Set QoS to low level of performance for file servers, virtual disks and LUNs used for backups, the performance of which will hardly be improved through automated QoS.

Replica Defrag: After the replica defrag is enabled, a replica will be aggregated on one node, and the business load will be preferentially redirected to the aggregate node, enabling local data read and reducing network overhead (the replica of stretched cluster 2 does not take effect). Recommended select Adaptive in most cases. adaptive mode adopt the most appropriate policy according to whether the mode is all-flash datastore or hybrid datastore. In the all-flash datastore mode, replica defrag is enabled; in the hybrid datastore mode, whether to aggregate or not is automatically selected according to the capacity of the cache disk and the business load.For a large-capacity database, if its capacity exceeds cache disk capacity of a single node, disable replica defrag to have replicas scattered across multiple nodes, so that cache disks on more than one node can be used, which improves cache disk hit rate and performance.

Default policy: You can set a storage policy as the default policy which will be preferentially used in scenarios such as creating and importing VMs, migrating VMs within a cluster, backing up and recovering VMs, creating shared virtual disks and LUNs.

A storage policy that is often used can be set as default to reduce repeated operations in scenarios such as creating and importing VMs, migrating VMs within a cluster, backing up and recovering VMs, creating shared disks and LUNs.