Skip to main content

Install with AeroLab

AeroLab creates Aerospike development and testing clusters in Docker, AWS, or GCP. It streamlines efforts to test cluster configuration options, upgrade procedures, and client applications in a controlled development environment.


AeroLab 7.5.0 or later is required for installing AVS. To install AeroLab, go to the latest release.

1. Generate an Aerospike configuration file

When installed with AeroLab, AVS expects the proximus-meta namespace to be in place. To generate a file with the full configuration, run the following:

aerolab conf generate

Select the vector checkbox and optionally the on-disk checkbox for the vector namespace. Press CTRL + X to save the configuration as aerospike.conf.

2. Verify AVS feature-key

Ensure that your feature-key file features.conf is in /etc/aerospike/ and includes the vector-service feature, as shown in the following example. Without that service, AVS will not start.

# generated 2024-02-23 19:31:59

feature-key-version 2
serial-number 12345

account-name Your Company
account-ID vector-trial

valid-until-date 2025-01-15

asdb-change-notification true
asdb-cluster-nodes-limit 0
asdb-compression true
asdb-encryption-at-rest true
asdb-flash-index true
asdb-ldap true
asdb-pmem true
asdb-rack-aware true
asdb-secrets true
asdb-strong-consistency true
asdb-vault true
asdb-xdr true
database-recovery true
elasticsearch-connector true
graph-service true
mesg-jms-connector true
mesg-kafka-connector true
presto-connector true
pulsar-connector true
spark-connector true
vector-service true

----- SIGNATURE ------------------------------------------------
MEY... ...mJA==
----- END OF SIGNATURE -----------------------------------------

To set your feature-key file in Aerolab, run the following:

aerolab config defaults -k '*.FeaturesFilePath' -v /etc/aerospike/features.conf

3. Create an Aerospike cluster

The following command creates a two-node cluster with a GCP-specific region and instance type this is ignored when deploying to Docker.

aerolab cluster create -n vectordb -c 2  -o aerospike.conf --zone us-central1-a --instance e2-standard-4

4. Create an AVS client machine

To create an AeroLab client for AVS, run the following:

aerolab client create vector -n vector -C vectordb --confirm --zone us-central1-a --instance e2-standard-4

The following command-line options apply specifically to your AVS cluster.

-C, --cluster-name=        aerospike cluster name to seed from (default: mydc)
--seed= specify an aerospike cluster seed IP:PORT instead of providing a ClusterName; if this parameter is provided, ClusterName is ignored
--listen= specify a listen IP:PORT for the service (default:
--no-touch-listen set this to prevent aerolab from touching the service: configuration part
--no-touch-seed set this to prevent aerolab from configuring the aerospike seed ip and port
--no-touch-advertised set this to prevent aerolab from configuring the advertised listeners
--version= vector version to install; only 0.3.1 is officially supported by aerolab (0.3.1-1 for rpm) (default: 0.3.1)
--custom-conf= provide a custom aerospike-proximus.yml to ship
--no-start if set, service will not be started after installation
-f, --featurefile= Features file to install; if not provided, the features.conf from the seed aerospike cluster will be taken
--metans= configure the metadata namespace name (default: proximus-meta)
--confirm set this parameter to confirm any warning questions without being asked to press ENTER to continue

5. Install example apps

The AVS client is best paired with any of the examples apps available from Aerospike. Follow these instructions to install example apps in the /opt directory.

  1. Clone the example apps repository.

    aerolab attach client -n vector
    apt -y install python3 python3-pip git
    cd /opt && git clone
  2. Install the prism image search app.

    aerolab client attach -n vector -- /bin/bash /opt/ install
  3. When the installation is complete, upload images:

    aerolab files upload -c -n vector {path-to-pictures-file-or-directory} /opt/proximus-examples/prism-image-search/prism/static/images/data/
  4. After all of the images have been uploaded, run the web server:

    aerolab client attach -n vector -- /bin/bash /opt/

    You now are ready to use AVS to perform an image search.

    • On Docker, go to
    • On cloud deployments, run aerolab client list, then go to http://EXTERNAL_IP:8080.