Kinetica Database Configuration¶

Using kubetctl a CustomResource of type KineticaCluster is used to define a new Kinetica DB Cluster in a yaml file.

The basic Group, Version, Kind or GVK to instantiate a Kinetica DB Cluster is as follows: -

1
2

apiVersion: app.kinetica.com/v1
kind: KineticaCluster

Metadata¶

to which we add a metadata: block for the name of the DB CR along with the namespace into which we are targetting the installation of the DB cluster.

1
2
3
4
5
6

apiVersion: app.kinetica.com/v1
kind: KineticaCluster
metadata:
  name: my-kinetica-db-cr
  namespace: gpudb
spec:

Spec¶

Under the spec: section of the KineticaCLuster CR we have a number of sections supporting different aspects of the deployed DB cluster:-

• gpudbCluster
• autoSuspend
• gadmin

gpudbCluster¶

Configuartion items specific to the DB itself.

1
2
3
4
5
6
7

apiVersion: app.kinetica.com/v1
kind: KineticaCluster
metadata:
  name: my-kinetica-db-cr
  namespace: gpudb
spec:
  gpudbCluster:

gpudbCluster¶

1
2
3
4
5
6
7

clusterName: kinetica-cluster 
clusterSize: 
  tshirtSize: M 
  tshirtType: LargeCPU 
fqdn: kinetica-cluster.saas.kinetica.com
haRingName: default
hasPools: false    

1. clusterName - the user defined name of the Kinetica DB Cluster

2. clusterSize - block that defines the number of DB Ranks to run

3. tshirtSize - sets the cluster size to a defined size based upon the t-shirt size. Valid sizes are: -

• XS - 1 DB Rank
• S - 2 DB Ranks
• M - 4 DB Ranks
• L - 8 DB Ranks
• XL - 16 DB Ranks
• XXL - 32 DB Ranks
• XXXL - 64 DB Ranks

4. tshirtType - block that defines the tyoe DB Ranks to run: -

• SmallCPU -
• LargeCPU -
• SmallGPU -
• LargeGPU -

5. fqdn - The fully qualified URL for the DB cluster. Used on the Ingress records for any exposed services.

6. haRingName - Default: default

7. hasPools - Whether to enable the separate node 'pools' for "infra", "compute" pod scheduling. Default: false +optional

autoSuspend¶

The DB Cluster autosuspend section allows for the spinning down of the core DB Pods to release the underlying Kubernetes nodes to reduce infrastructure costs when the DB is not in use.

1
2
3
4
5
6
7
8
9

apiVersion: app.kinetica.com/v1
kind: KineticaCluster
metadata:
  name: my-kinetica-db-cr
  namespace: gpudb
spec:
  autoSuspend:
    enabled: false
    inactivityDuration: 1h0m0s

7. the start of the autoSuspend definition

8. enabled when set to true auto suspend of the DB cluster is enabled otherwise set to false and no automatic suspending of the DB takes place. If omitted it defaults to false

9. inactivityDuration the duration after which if no DB activity has taken place the DB will be suspended

gadmin¶

GAdmin the Database Administration Console

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12

apiVersion: app.kinetica.com/v1
kind: KineticaCluster
metadata:
  name: my-kinetica-db-cr
  namespace: gpudb
spec:
  gadmin:
    containerPort:
      containerPort: 8080
      name: gadmin
      protocol: TCP
    isEnabled: true

7. gadmin configuration block definition

8. containerPort configuration block i.e. where gadmin is exposed on the DB Pod

9. containerPort the port number as an integer. Default: 8080

10. name the name of the port being exposed. Default: gadmin

11. protocol network protocal used. Default: TCP

12. isEnabled whether gadmin is exposed from the DB pod. Default: true