SingleStore Managed Service

Create a Managed Service Cluster

As SingleStore Managed Service manages its own clusters, customers can create new clusters without the need to configure or physically manage them.

Customers can create and use a single trial cluster at a time, which automatically terminates after eight (8) hours have elapsed from the time the cluster was created. The free trial cluster is limited to a development cluster with four (4) SingleStore DB units, for a total of 32 vCPUs, 256GB of RAM, and 4 TBs of storage.

While trial clusters continue to run when you log out, there are no cost implications by leaving your cluster online.

There are a few options available when creating a SingleStore Managed Service cluster, including:

Region

Regions are datacenter geographies in which a SingleStore Managed Service cluster runs. While any region can be selected, select the region that corresponds to the same region in which your target application is running. Note that regions vary in price.

Cluster Size

Cluster sizing is the same as with self-managed SingleStore DB deployments. Consideration should be given to compute, memory, and storage needs when selecting cluster size. A cluster’s size can later be increased or decreased as required.

Cluster Restriction

SingleStore Managed Service provides for Classless Inter-Domain Routing (CIDR) whitelisting to restrict the networks that can access a cluster. This can be defined when initially creating the cluster, or modified via editing the IP Whitelist once the cluster has been created.

Columnstore as Default

The default_table_type engine variable specifies whether CREATE TABLE creates a default rowstore or columnstore table. In SingleStore Managed Service, the default is set to columnstore for newly created v7.3 clusters only. For existing organizations, this variable remains as rowstore when creating new clusters.

While the columnstore-as-default feature now creates columnstore tables by default via the standard CREATE TABLE syntax, you may also create a rowstore table explicitly via the CREATE ROWSTORE TABLE ... syntax, regardless of the setting of default_table_type.

Associated Changes
Hash Keys and Sort Keys

A standard CREATE TABLE statement using KEY(k) stores <k> as a hash key in the resulting columnstore table.

Hash keys support fast equality lookups, but not fast range filters. To create an index to help support fast range filters on a columnstore table, a sort key <k> can be used by explicitly specifying KEY(k) USING CLUSTERED COLUMNSTORE in your CREATE TABLE statement.

Ordered Index

Using more than one ordered index is no longer supported. Effectively, a table may only have one, which is the columnstore sort key.

Multi-column Unique Keys

Using multi-column unique keys in a columnstore table is supported starting in engine version 7.5.

Composite Keys

Lookups of a composite key may now take longer. For example, instead of seeking on one index KEY(a,b), the KEY(a) index and the KEY(b) index must both be searched, and the key sets must then be intersected to determine which rows to look up.

Additional Information

Refer to Creating a Columnstore Table Using Standard CREATE TABLE Syntax for additional information.