Activate YTsaurus

YTsaurus deploys a single shared YTsaurus cluster on the platform. The cluster is described by a YTsaurusConfig resource named main. By default, the component is disabled (spec.enabled: false); to deploy the cluster, set the resource to spec.enabled: true.

Using the CLIUsing the CLI

1. Open the current YTsaurusConfig main resource to edit it: kubectl edit ytsaurusconfig main. If you prefer to apply the configuration from a file, upload the resource kubectl get ytsaurusconfig main -o yaml > ytsaurusconfig.yaml.

2. Replace the content with one of the configurations below:

   Minimum configuration

   Extended configuration

   Includes a minimal set of subcomponents: cluster core, web interface, and HTTP proxy. Suitable for a test environment.

   apiVersion: stackland.yandex.cloud/v1alpha1
   kind: YTsaurusConfig
   metadata:
     name: main
   spec:
     enabled: true
     clusterName: cluster
     cluster:
       discovery:
         instanceCount: 1
       primaryMasters:
         instanceCount: 1
         storage:
           size: "5Gi"
       httpProxies:
         instanceCount: 1
       dataNodes:
         - name: ssd
           instanceCount: 3
           storage:
             size: "20Gi"
       execNodes:
         - name: default
           instanceCount: 1
           storage:
             size: "10Gi"
       schedulers:
         instanceCount: 1
       controllerAgents:
         instanceCount: 1
       bundleController:
         enabled: true
       queryTrackers:
         instanceCount: 1
       queueAgents:
         instanceCount: 1
       rpcProxies:
         instanceCount: 1
       strawberry:
         enabled: true
       yqlAgents:
         instanceCount: 1
       ui:
         enabled: true
         ingressEnabled: true
   

   A full set of subcomponents with explicitly specified resources and storage: tablet nodes, Strawberry, CHYT, SPYT, Task Proxy, Cron. Suitable for a production environment.

   apiVersion: stackland.yandex.cloud/v1alpha1
   kind: YTsaurusConfig
   metadata:
     name: main
   spec:
     enabled: true
     clusterName: cluster
     cluster:
       discovery:
         instanceCount: 1
         resources:
           cpu: "100m"
           memory: "256Mi"
   
       primaryMasters:
         instanceCount: 1
         storage:
           storageClass: stackland-ssd
           size: "10Gi"
         resources:
           cpu: "500m"
           memory: "1Gi"
   
       httpProxies:
         instanceCount: 2
         resources:
           cpu: "500m"
           memory: "1Gi"
   
       dataNodes:
         - name: ssd
           instanceCount: 3
           storage:
             storageClass: stackland-ssd
             size: "50Gi"
           resources:
             cpu: "500m"
             memory: "1Gi"
   
       execNodes:
         - name: default
           instanceCount: 2
           storage:
             storageClass: stackland-ssd
             size: "20Gi"
           resources:
             cpu: "4000m"
             memory: "4Gi"
   
       schedulers:
         instanceCount: 1
         resources:
           cpu: "500m"
           memory: "512Mi"
   
       controllerAgents:
         instanceCount: 1
         resources:
           cpu: "500m"
           memory: "512Mi"
   
       bundleController:
         enabled: true
   
       queryTrackers:
         instanceCount: 1
         resources:
           cpu: "200m"
           memory: "512Mi"
   
       queueAgents:
         instanceCount: 1
         resources:
           cpu: "200m"
           memory: "512Mi"
   
       rpcProxies:
         instanceCount: 2
         resources:
           cpu: "200m"
           memory: "512Mi"
   
       strawberry:
         enabled: true
         resources:
           cpu: "200m"
           memory: "512Mi"
   
       tabletNodes:
         - name: default
           instanceCount: 2
           resources:
             cpu: "1000m"
             memory: "2Gi"
   
       yqlAgents:
         instanceCount: 1
         resources:
           cpu: "500m"
           memory: "1Gi"
   
       ui:
         enabled: true
         ingressEnabled: true
         resources:
           cpu: "200m"
           memory: "512Mi"
   
       cron:
         clear_tmp:
           enabled: true
   
       taskProxy:
         enabled: true
         instanceCount: 1
   
       spyt:
         enabled: true
   
       chyt:
         enabled: true
   
       tutorial:
         enabled: false
   
       odin:
         enabled: false
   

3. Apply the manifest: kubectl apply -f ytsaurusconfig.yaml. The resource operates at the cluster level; you do not need to specify a namespace.

4. Wait for the component to enter the ready state. To check the installatioln status, run this command:

   kubectl get componentinstallation ytsaurus-main -o jsonpath='{.status.phase}'
   

   The possible values are:

   • Installing: Installing the component.
   • Updating: Applying the new configuration.
   • Ready: Component installed and ready for operation.
   • Error: Error. For details, refer to the status.message field.

5. You can check the YTsaurus cluster status using the following command:

   kubectl -n stackland-ytsaurus get ytsaurus cluster
   

   The resource name corresponds to spec.clusterName in YTsaurusConfig. Wait until the cluster turns Running. Its initial start may take 10–20 minutes.

Next stepsNext steps

After you enable the component:

• Get your administrator password and token.
• Open the YTsaurus web interface.
• Connect to the YTsaurus HTTP proxy.
• Update the cluster settings if you need to edit the resources, node group composition, or subcomponents.