Delete OBServer nodes from zones

This topic describes how to scale in an OceanBase cluster by deleting OBServer nodes from zones.

Make sure that the zones still have sufficient resources to bear the load of the tenant after you delete the OBServer nodes.

You can modify the obcluster.yaml configuration file of the OceanBase cluster to delete OBServer nodes from zones. Generally, you must delete the same number of OBServer nodes from each zone.

The spec.topology.cluster[index1].zone[index2].replicas attribute in the obcluster.yaml configuration file specifies the number of OBServer nodes in the [index2] zone of the OceanBase cluster deployed in the [index1] Kubernetes cluster. You can modify the value of this attribute to delete OBServer nodes from a zone.

Procedure

Assume that the current cluster has three zones, zone1, zone2, and zone3, each zone contains two OBServer nodes, and you want to delete one OBServer node from each zone for scale-in. After the scale-in, each of the three zones has only one OBServer node.

1. Modify the obcluster.yaml configuration file to change the value of replicas from 2 to 1 for each zone.

   # For example, each zone has two OBServer nodes.
   spec:
     topology:
       - cluster: cn
         zone:
         - name: zone1
           region: region1
           nodeSelector:
             ob.zone: zone1
           replicas: 2
         - name: zone2
           region: region1
           nodeSelector:
             ob.zone: zone2
           replicas: 2
         - name: zone3
           region: region1
           nodeSelector:
             ob.zone: zone3
           replicas: 2
   
   # Reduce the number of OBServer nodes in each zone to 1.
   spec:
     topology:
       - cluster: cn
         zone:
         - name: zone1
           region: region1
           nodeSelector:
             ob.zone: zone1
           replicas: 1 # 2 -> 1
         - name: zone2
           region: region1
           nodeSelector:
             ob.zone: zone2
           replicas: 1 # 2 -> 1
         - name: zone3
           region: region1
           nodeSelector:
             ob.zone: zone3
           replicas: 1 # 2 -> 1
   

2. Run the following command for the modification to take effect:

   kubectl apply -f obcluster.yaml
   

3. Run the following command to check whether the modification has taken effect:

   kubectl get pod -n obcluster
   

   Here is an example of the command output:

   NAME                      READY   STATUS    RESTARTS   AGE
   sapp-ob-test-cn-zone1-0   2/2     Running   0          84m
   sapp-ob-test-cn-zone3-0   2/2     Running   0          84m
   sapp-ob-test-cn-zone2-0   2/2     Running   0          84m
   

   The preceding result shows that zone1, zone2, and zone3 each have only one OBServer node, the values of READY are all 2/2, and the values of STATUS are all Running, indicating that the scale-in is successful.