Skip to main content
Version: release-0.9

Manage Qdrant with KubeBlocks

The popularity of generative AI (Generative AI) has aroused widespread attention and completely ignited the vector database (Vector Database) market. KubeBlocks supports the management of vector databases, such as Qdrant, Milvus, and Weaviate.

In this chapter, we take Qdrant as an example to show how to manage vector databases with KubeBlocks. This chapter illustrates how to create and manage a Qdrant cluster by kubectl or a YAML file. You can find the YAML examples in the GitHub repository.

Before you start

  • Install KubeBlocks.

  • View all the database types and versions available for creating a cluster.

    Make sure the qdrant cluster definition is installed. If the cluster definition is not available, refer to this doc to enable it first.

    kubectl get clusterdefinition qdrant
    >
    NAME TOPOLOGIES SERVICEREFS STATUS AGE
    qdrant Available 30m

    View all available versions for creating a cluster.

    kubectl get clusterversions -l clusterdefinition.kubeblocks.io/name=qdrant
  • To keep things isolated, create a separate namespace called demo throughout this tutorial.

    kubectl create namespace demo

Create a cluster

KubeBlocks implements a Cluster CRD to define a cluster. Here is an example of creating a Qdrant Replication cluster. Primary and Secondary are distributed on different nodes by default. But if you only have one node for deploying a Replication Cluster, set spec.affinity.topologyKeys as null.

cat <<EOF | kubectl apply -f -
apiVersion: apps.kubeblocks.io/v1alpha1
kind: Cluster
metadata:
name: mycluster
namespace: demo
spec:
clusterDefinitionRef: qdrant
clusterVersionRef: qdrant-1.8.1
terminationPolicy: Delete
affinity:
podAntiAffinity: Preferred
topologyKeys:
- kubernetes.io/hostname
tolerations:
- key: kb-data
operator: Equal
value: 'true'
effect: NoSchedule
componentSpecs:
- name: qdrant
componentDefRef: qdrant
disableExporter: true
serviceAccountName: kb-mycluster
replicas: 2
resources:
limits:
cpu: '0.5'
memory: 0.5Gi
requests:
cpu: '0.5'
memory: 0.5Gi
volumeClaimTemplates:
- name: data
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 20Gi
EOF
FieldDefinition
spec.clusterDefinitionRefIt specifies the name of the ClusterDefinition for creating a specific type of cluster.
spec.clusterVersionRefIt is the name of the cluster version CRD that defines the cluster version.
spec.terminationPolicyIt is the policy of cluster termination. The default value is Delete. Valid values are DoNotTerminate, Halt, Delete, WipeOut.

- DoNotTerminate blocks deletion operation.

- Halt deletes workload resources such as statefulset and deployment workloads but keep PVCs.

- Delete is based on Halt and deletes PVCs.

- WipeOut is based on Delete and wipe out all volume snapshots and snapshot data from a backup storage location.
spec.affinityIt defines a set of node affinity scheduling rules for the cluster's Pods. This field helps control the placement of Pods on nodes within the cluster.
spec.affinity.podAntiAffinityIt specifies the anti-affinity level of Pods within a component. It determines how pods should spread across nodes to improve availability and performance.
spec.affinity.topologyKeysIt represents the key of node labels used to define the topology domain for Pod anti-affinity and Pod spread constraints.
spec.tolerationsIt is an array that specifies tolerations attached to the cluster's Pods, allowing them to be scheduled onto nodes with matching taints.
spec.componentSpecsIt is the list of components that define the cluster components. This field allows customized configuration of each component within a cluster.
spec.componentSpecs.componentDefRefIt is the name of the component definition that is defined in the cluster definition and you can get the component definition names with kubectl get clusterdefinition apecloud-mysql -o json \| jq '.spec.componentDefs[].name'.
spec.componentSpecs.nameIt specifies the name of the component.
spec.componentSpecs.disableExporterIt defines whether the monitoring function is enabled.
spec.componentSpecs.replicasIt specifies the number of replicas of the component.
spec.componentSpecs.resourcesIt specifies the resource requirements of the component.

KubeBlocks operator watches for the Cluster CRD and creates the cluster and all dependent resources. You can get all the resources created by the cluster with kubectl get all,secret,rolebinding,serviceaccount -l app.kubernetes.io/instance=mycluster -n demo.

kubectl get all,secret,rolebinding,serviceaccount -l app.kubernetes.io/instance=mycluster -n demo

Run the following command to see the created Qdrant cluster object:

kubectl get cluster mycluster -n demo -o yaml

Connect to a vector database cluster

Qdrant provides both HTTP and gRPC protocols for client access on ports 6333 and 6334 respectively. You can also port forward the service to connect to a database from your local machine.

  1. Run the following command to port forward the service.

    kubectl port-forward svc/mycluster-qdrant 6333:6333 -n demo
  2. Open a new terminal and run the following command to connect to the database.

    curl http://127.0.0.1:6333/collections

    Refer to the official Qdrant documents for the cluster operations.

Scaling

Scaling function for vector databases is also supported.

Scale horizontally

Horizontal scaling changes the amount of pods. For example, you can apply horizontal scaling to scale pods up from three to five. The scaling process includes the backup and restore of data.

Before you start

Check whether the cluster STATUS is Running. Otherwise, the following operations may fail.

kubectl get cluster mycluster -n demo
>
NAME CLUSTER-DEFINITION VERSION TERMINATION-POLICY STATUS AGE
mycluster qdrant qdrant-1.5.0 Delete Running 47m

Steps

There are two ways to apply horizontal scaling.

  1. Apply an OpsRequest to a specified cluster. Configure the parameters according to your needs.

    kubectl apply -f - <<EOF
    apiVersion: apps.kubeblocks.io/v1alpha1
    kind: OpsRequest
    metadata:
    name: ops-horizontal-scaling
    namespace: demo
    spec:
    clusterName: mycluster
    type: HorizontalScaling
    horizontalScaling:
    - componentName: qdrant
    replicas: 1
    EOF
  2. Check the operation status to validate the horizontal scaling status.

    kubectl get ops -n demo
    >
    NAMESPACE NAME TYPE CLUSTER STATUS PROGRESS AGE
    demo ops-horizontal-scaling HorizontalScaling mycluster Succeed 3/3 6m

    If an error occurs to the horizontal scaling operation, you can troubleshoot with kubectl describe ops -n demo command to view the events of this operation.

  3. Check whether the corresponding resources change.

    kubectl describe cluster mycluster -n demo

Handle the snapshot exception

If STATUS=ConditionsError occurs during the horizontal scaling process, you can find the cause from cluster.status.condition.message for troubleshooting. In the example below, a snapshot exception occurs.

Status:
conditions:
- lastTransitionTime: "2024-04-25T17:40:26Z"
message: VolumeSnapshot/mycluster-qdrant-scaling-dbqgp: Failed to set default snapshot
class with error cannot find default snapshot class
reason: ApplyResourcesFailed
status: "False"
type: ApplyResources

Reason

This exception occurs because the VolumeSnapshotClass is not configured. This exception can be fixed after configuring VolumeSnapshotClass, but the horizontal scaling cannot continue to run. It is because the wrong backup (volumesnapshot is generated by backup) and volumesnapshot generated before still exist. First, delete these two wrong resources and then KubeBlocks re-generates new resources.

Steps:

  1. Configure the VolumeSnapshotClass by running the command below.

    kubectl create -f - <<EOF
    apiVersion: snapshot.storage.k8s.io/v1
    kind: VolumeSnapshotClass
    metadata:
    name: csi-aws-vsc
    annotations:
    snapshot.storage.kubernetes.io/is-default-class: "true"
    driver: ebs.csi.aws.com
    deletionPolicy: Delete
    EOF
  2. Delete the wrong backup (volumesnapshot is generated by backup) and volumesnapshot resources.

    kubectl delete backup -l app.kubernetes.io/instance=mycluster

    kubectl delete volumesnapshot -l app.kubernetes.io/instance=mycluster

Result

The horizontal scaling continues after backup and volumesnapshot are deleted and the cluster restores to running status.

Scale vertically

You can vertically scale a cluster by changing resource requirements and limits (CPU and storage). For example, if you need to change the resource class from 1C2G to 2C4G, vertical scaling is what you need.

note

During the vertical scaling process, all pods restart and the pod role may change after the restarting.

Before you start

Check whether the cluster status is Running. Otherwise, the following operations may fail.

kubectl get cluster mycluster -n demo
>
NAME CLUSTER-DEFINITION VERSION TERMINATION-POLICY STATUS AGE
mycluster qdrant qdrant-1.5.0 Delete Running 47m

Steps

There are two ways to apply vertical scaling.

  1. Apply an OpsRequest to the specified cluster. Configure the parameters according to your needs.

    kubectl apply -f - <<EOF
    apiVersion: apps.kubeblocks.io/v1alpha1
    kind: OpsRequest
    metadata:
    name: ops-vertical-scaling
    namespace: demo
    spec:
    clusterName: mycluster
    type: VerticalScaling
    verticalScaling:
    - componentName: qdrant
    requests:
    memory: "2Gi"
    cpu: "1"
    limits:
    memory: "4Gi"
    cpu: "2"
    EOF
  2. Check the operation status to validate the vertical scaling.

    kubectl get ops -n demo
    >
    NAMESPACE NAME TYPE CLUSTER STATUS PROGRESS AGE
    demo ops-vertical-scaling VerticalScaling mycluster Succeed 3/3 6m

    If an error occurs to the vertical scaling operation, you can troubleshoot with kubectl describe ops -n demo command to view the events of this operation.

  3. Check whether the corresponding resources change.

    kubectl describe cluster mycluster -n demo

Volume Expanding

Before you start

Check whether the cluster status is Running. Otherwise, the following operations may fail.

kubectl get cluster mycluster -n demo
>
NAME CLUSTER-DEFINITION VERSION TERMINATION-POLICY STATUS AGE
mycluster qdrant qdrant-1.5.0 Delete Running 4m29s

Steps

There are two ways to apply volume expansion.

  1. Change the value of storage according to your need and run the command below to expand the volume of a cluster.

    kubectl apply -f - <<EOF
    apiVersion: apps.kubeblocks.io/v1alpha1
    kind: OpsRequest
    metadata:
    name: ops-volume-expansion
    namespace: demo
    spec:
    clusterName: mycluster
    type: VolumeExpansion
    volumeExpansion:
    - componentName: qdrant
    volumeClaimTemplates:
    - name: data
    storage: "40Gi"
    EOF
  2. Validate the volume expansion operation.

    kubectl get ops -n demo
    >
    NAMESPACE NAME TYPE CLUSTER STATUS PROGRESS AGE
    demo ops-volume-expansion VolumeExpansion mycluster Succeed 3/3 6m

    If an error occurs to the vertical scaling operation, you can troubleshoot with kubectl describe ops -n demo command to view the events of this operation.

  3. Check whether the corresponding cluster resources change.

    kubectl describe cluster mycluster -n demo

Stop/Start a cluster

You can stop/start a cluster to save computing resources. When a cluster is stopped, the computing resources of this cluster are released, which means the pods of Kubernetes are released, but the storage resources are reserved. Start this cluster again if you want to restore the cluster resources from the original storage by snapshots.

Stop a cluster

Run the command below to stop a cluster.

kubectl apply -f - <<EOF
apiVersion: apps.kubeblocks.io/v1alpha1
kind: OpsRequest
metadata:
name: ops-stop
namespace: demo
spec:
clusterName: mycluster
type: Stop
EOF

Start a cluster

Run the command below to start a cluster.

kubectl apply -f - <<EOF
apiVersion: apps.kubeblocks.io/v1alpha1
kind: OpsRequest
metadata:
name: ops-start
namespace: demo
spec:
clusterName: mycluster
type: Start
EOF

Backup and restore

The backup and restore operations for Qdrant are the same with those of other clusters.

You can refer to the backup and restore docs for details.