Aerospike Vector Search

This Helm chart allows you to configure and run our official Aerospike Vector Search docker image on a Kubernetes cluster.

This helm chart sets up a StatefulSet for each AVS instance. We use a StatefulSet instead of a Deployment, to have stable DNS names for the
deployed AVS pods.

Prerequisites

• Kubernetes cluster
• Helm v3
• An Aerospike cluster that can connect to Pods in the Kubernetes cluster. The Aerospike cluster can be deployed in the same Kubernetes cluster using Aerospike Kubernetes Operator
• Ability to deploy a LoadBalancer on K8s in case AVS app runs outside the Kubernetes cluster

Adding the helm chart repository

Add the aerospike-helm helm repository if not already done. (Note: The repository has moved to artifact.aerospike.io. If you are still pointing to aerospike.github.io, please update the repository URL)

helm repo add aerospike-io https://artifact.aerospike.io/artifactory/api/helm/aerospike-helm

Supported configuration

Configuration

Parameter	Description	Default
replicaCount	Configures the number AVS instance pods to run.	'1'
image	Configures AVS image repository, tag and pull policy.	see values.yaml
imagePullSecrets	For Private docker registries, when authentication is needed.	see values.yaml
aerospikeVectorSearchConfig	AVS cluster configuration deployed to /etc/aerospike-vector-search/aerospike-vector-search.yml.	see values.yaml
initContainers	List of initContainers added to each AVS pods for custom cluster behavior.	[]
serviceAccount	Service Account details like name and annotations.	see values.yaml
podAnnotations	Additional pod annotations. Should be specified as a map of annotation names to annotation values.	{}
podLabels	Additional pod labels. Should be specified as a map of label names to label values.	{}
podSecurityContext	Pod security context	{}
securityContext	Container security context	{}
service	Load-Balancer configuration for more details please refer to a Load-Balancer docs.	{}
resources	Resource requests and limits for the AVS pods.	{}
autoscaling	Enable the horizontal pod auto-scaler.	see values.yaml
extraVolumes	List of additional volumes to attach to the AVS pod.	see values.yaml
extraVolumeMounts	Extra volume mounts corresponding to the volumes added to extraVolumes.	see values.yaml
extraSecretVolumeMounts	Extra secret volume mounts corresponding to the volumes added to extraVolumes.	see values.yaml
affinity	Affinity rules if any for the pods.	{}
nodeSelector	Node selector for the pods.	{}
tolerations	Tolerations for the pods.	{}

Deploy the AVS Cluster

We recommend creating a new .yaml for providing configuration values to the helm chart for deployment. See the Aerospike Vector repository for examples and deployment scripts.

A sample values yaml file is shown below:

replicaCount: 1

image:
  tag: "1.0.0"

aerospikeVectorSearchConfig:
  aerospike:
    metadata-namespace: "avs-meta"
    seeds:
      - aerospike-cluster-0-0.aerospike-cluster.aerospike.svc.cluster.local:
          port: 3000

Here replicaCount is the count of AVS pods that are deployed. The AVS configuration is provided as yaml under the key aerospikeVectorSearchConfig.

See Aerospike Vector Search configuration documentation for more details on AVS configuration.

Create a new namespace

We recommend using a namespace for the AVS cluster. If the namespace does not exist run the following command:

kubectl create namespace avs

Create secrets

Create the secret for avs using your Aerospike licence file

# kubectl --namespace <target namespace> create secret generic avs-secret--from-file=features.conf=<path to features conf file>
kubectl --namespace avs create secret generic aerospike-secret --from-file=features.conf=features.conf

Deploy the AVS cluster

# helm install --namespace <target namespace> <helm release name/cluster name> -f <path to custom values yaml> aerospike/aerospike-vector-search
helm install  avs-chart --namespace avs -f avs-values.yaml aerospike/aerospike-vector-search

Here avs is the release name for the AVS cluster and also its cluster name.

On successful deployment you should see output similar to below:

NAME: avs
LAST DEPLOYED: Tue May 21 15:55:39 2024
NAMESPACE: avs
STATUS: deployed
REVISION: 1
TEST SUITE: None
NOTES:

List pods for the AVS cluster

To list the pods for the AVS cluster run the following command:

# kubectl get pods --namespace aerospike --selector=app=<helm release name>-aerospike-vector-search
kubectl get pods --namespace aerospike --selector=app=avs-aerospike-vector-search

You should see output similar to the following:

NAME                            READY   STATUS    RESTARTS   AGE
avs-aerospike-vector-search-0   1/1     Running   0          9m52s

If you are using Aerospike Kubernetes Operator, see quote-search for reference.

Get logs for all AVS instances

# kubectl -n aerospike logs -f statefulset/<helm release name>-aerospike-vector-search
# Skip the -f flag to get a one time dump of the log
kubectl -n aerospike logs -f statefulsets/avs-aerospike-vector-search

Get logs for one AVS pod

# kubectl -n aerospike logs -f <helm release name>-aerospike-vector-search-0
# Skip the -f flag to get a one time dump of the log
kubectl -n aerospike logs -f avs-aerospike-vector-search-0

Updating AVS configuration

Edit the aerospikeVectorSearchConfig section in the custom values file and save the changes.

Upgrade the AVS deployment using the following command.

#helm upgrade --namespace <target namespace> <helm release name> -f <path to custom values yaml file> aerospike/aerospike-vector-search
helm upgrade --namespace avs avs-chart -f avs-values.yaml aerospike/aerospike-vector-search

On successful execution of the command the AVS pods will undergo a rolling restart and come up with the new configuration.

To verify the changes are applied

• List the pods
• Verify the configuration in AVS logs

NOTE: The changes might take some time to apply. If you do not see the desired AVS config try again after some time.

Scaling up/down the AVS instances

Edit the replicaCount to the desired AVS instance count and upgrade the AVS deployment using the following command.

#helm upgrade --namespace <target namespace> <helm release name> -f <path to custom values yaml file> aerospike/aerospike-vector-search
helm upgrade --namespace avs avs-chart -f avs-values.yaml aerospike/aerospike-vector-search

Verify that the AVS cluster have been scaled.

• List the pods and verify the count of AVS instances is as desired

NOTE: The changes might take some time to apply. If you do not see the desired count try again after some time.