Manage data pools
Users use the block storage resources in the storage cluster through data pools. Data pools define the replication and erasure coding data redundancy policies.
Restrictions and guidelines
You must create data pools after the system is initialized. You can create a maximum of five data pools.
To select file storage disk pools for creating data pools, you must create file systems before you create data pools.
For important services and scenarios that require high reliability, set the number of replicas to 3 or more if the data redundancy policy is replication.
Add a data pool
On the top navigation bar, click Storage, and then select Storage Management > Data Pool Management from the navigation pane.
Select a cluster from the Cluster list in the upper right corner of the page if multiple clusters exist in the system.
Click Add.
Configure the parameters as described in "Parameters."
Click OK.
Edit a data pool
On the top navigation bar, click Storage, and then select Storage Management > Data Pool Management from the navigation pane.
Select a cluster from the Cluster list in the upper right corner of the page if multiple clusters exist in the system.
Click the
icon in the Actions column for the target data pool.
Set the minimum number of replicas.
By default, the minimum number of replicas is one less than the number of data pool replicas. The minimum value for the minimum number of replicas parameter is one, and the maximum value for the minimum number of replicas parameter is equivalent to the number of replicas.
Click OK.
Delete a data pool
On the top navigation bar, click Storage, and then select Storage Management > Data Pool Management from the navigation pane.
Select a cluster from the Cluster list in the upper right corner of the page if multiple clusters exist in the system.
Select a data pool, and then click Delete.
Click OK in the dialog box that opens.
Parameters
Add a data pool:
Cluster: Cluster to which the data pool belongs. This field is not configurable. You must select a cluster from the Cluster list in the upper right corner of the page before adding a data pool. The Cluster list is available only when multiple clusters exist in the system.
Disk Pool: Select a disk pool. The size and data storage type of the data pool depend on the disk pool you select.
Redundancy Policy: Select a redundancy policy type.
Replication: If you select this option, set the number of replicas. If the number of storage nodes is 5 or larger, the maximum number of replicas is 5. If the number of storage nodes is smaller than 5, the minimum number of replicas is 2, and the maximum number of replicas is equal to the number of storage nodes.
Erasure Coding: If you select this option, set the number of data blocks, number of parity blocks, and the stripe size.
|
|
To use ensure coding as the redundancy policy, first contact the technical support for evaluation. |
Replicas: Set the number of replicas. For important services and scenarios that require high reliability, use three or more replicas as a best practice. If the storage cluster has only two storage nodes, you must set the number of replicas to 2.
Data Blocks: Set the number of blocks into which data is divided. Each data block is stored on a disk of a storage node.
Parity Blocks: Set the number of parity blocks used in conjunction with data blocks to recode data. A parity block is saved on a disk of a storage node.
Strip Size: Set the size of data blocks used for encoding and decoding. A small strip size is suitable for random read/write. A large strip size is suitable for sequential read/write. As a best practice, set a large stripe size in VM scenarios and a small stripe size in database application scenarios.
Chunk Size—Set the chunk size for the data blocks. To ensure load balancing between nodes, data will be divided into multiple blocks in the same size. To ensure service performance, set the chunk size according to actual service conditions.
Mount as UIS RBD Storage Pool: Select whether to mount the data pool as a UIS RBD storage pool. This parameter is available only when a thin-provisioned disk pool is selected.
Data pool list:
Policy Details: When the policy type is Replication, this field displays the number of replicas. When the policy type is Erasure Coding, this field displays the number of data blocks, the number of parity blocks, and the stripe size in number of data blocks + number parity blocks + stripe size format, 2 + 1+ 8k, for example. Each of the two data blocks and one parity block of one piece of data will be stored on a disk of three different storage nodes. The data is available when one of the three disks fails, and the storage usage is 66.67%.
Total Capacity: Total capacity of the data pool.
Assigned Capacity: Assigned capacity of the data pool.
Available Capacity: Assignable capacity of the data pool.
State: State of the data pool.
Normal: The data pool is operating correctly.
Abnormal: The disk pool does not meet the redundancy policy requirements of the data pool. For example, if the disk pool uses hosts as fault domains and the data pool needs three replicas, the disk pool must provide a minimum of three disks located on different hosts for the data pool. If the disk pool has only two available disks, the state of the data pool is abnormal.
Subhealthy: The disk pool meets the redundancy policy requirements of the data pool, but disk failures have occurred or significant data distribution skew exists.