How to run weaviate instance on a VM instead of Docker

Can someone provide the steps how to configure Weaviate on a VM and steps to configure HA

Hi @sri-postgres-cassand ! Welcome to our community :hugs:

While we still suggest running Weaviate using Docker or K8s, you can always run it directly from the binary.

for that, you can download the binary from our github page, for example:

I have just downloaded, for instance, as I am running on mac:
weaviate-v1.23.0-darwin-all.zip

now, inside this zip you have Weaviate’s binary. You can make it executable:

chmod + x weaviate

and now run Weaviate directly thru the binary, defining the bare minimum, like so:

export AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED=true; 
export PERSISTENCE_DATA_PATH=/tmp/weaviate; 
./weaviate --host 0.0.0.0 --port '8080' --scheme http

Let me know if this helps :slight_smile:

Thanks!

Thank you Very Much @DudaNogueira . This helped to configure a Weaviate standalone instance on a VM.

I think it would be better if there is a way to configure Replication and Horizontal Scaling on VM along with Docker and K8s.

Hi!

For K8s you can use our helm. Adding replicas is a matter of changing the values and deployig:

For docker, you can generate your docker compose using the configurator tool:

Then adapt it to spin multiple nodes, changing the necessary env variables.

:slight_smile:

Any particular reason for recommending docker over running the binary directly, if they’re both deployed as standalone?

Hi @emmastro !

Welcome to our community :hugs:

No particular reason.

In fact, for production, we suggest K8s, so you can have orchestration and other tools to help managing it that you will not get while running docker alone.

1 Like