Applies to VAST Cluster 5.5.
Before you can build on a service, your platform team sets a few things up on the cluster. Most of them are the same for every service. This page describes them once, in plain terms, and then links each service's own setup or access page, which says exactly what to ask for and gives you a request you can copy.
What every service needs
Admin task: send to your platform team. You don't set these up yourself. The procedures are in the KB, linked from each service's setup or access page.
| You need | What it is |
|---|---|
| An endpoint | The address (a DNS name or virtual IPs) of a virtual IP pool that serves the service and your tenant. Your client connects to it over the network. It isn't the management (VMS) address. |
| A VAST user | The identity your requests run as. A user belongs to one tenant, and the cluster checks that user's permissions on each request. |
| S3 access keys | An access key and secret key that belong to that user. Clients sign their requests with them, also when the service isn't object storage: the VAST DataBase SDK, Trino, Spark and the VAST Query Engine use them too. |
| An identity policy | A policy attached to the user (or its group) that grants the actions you need on the resources you'll use, for example querying one database or writing under one bucket prefix. A missing action shows up as a Forbidden (HTTP 403) error. |
You also need network access from your client, or from wherever your code runs, to the endpoint. Each service then adds something of its own, such as a database, a bucket or a broker; the table below lists it.
These pages have the details:
- Setup prerequisites (VAST DataBase) explains the endpoint, keys and identity policy, and which actions each SDK operation needs.
- Access to VAST S3 and NFS explains identity and bucket policies, and temporary credentials.
Each service's setup
Each setup or access page has a diagram of what your platform team sets up, in order, before your client connects. The same diagram is under "Before you start" in that service's quickstart.
| Service | What it adds to the common items | Setup or access page | Setup diagram |
|---|---|---|---|
| VAST DataBase | A database: a bucket on a Database-enabled view. For the VAST Query Engine, a dedicated Query Engine virtual IP pool and s3:QueryEngine*. |
Setup prerequisites | Checklist |
| VAST Vector Search | Everything VAST DataBase needs, plus a Query Engine endpoint (a virtual IP pool with the Query Engine role) and the ADBC driver. | Setup | What you need |
| VAST Event Broker | A Kafka view (your broker), bootstrap addresses from its virtual IP pool, and network access to every virtual IP in the pool. Kafka clients authenticate with no credentials, SASL/PLAIN (a VAST local user and password, not S3 keys) or mTLS; S3 keys are needed only to read topics as tables. | Setup | What you need |
| VAST DataEngine | DataEngine enabled on your tenant, an event broker, a compute cluster and namespace, a container registry, and a DataEngine user with a password. An S3 key pair for you, and one for your functions if they read or write objects or tables. | Setup prerequisites | What you need |
| VAST DataStore | A bucket (or an NFS export) on a view, and the view policy's security flavor, which decides whose permissions apply. | Access | What to ask for |
| VAST Catalog | The Catalog enabled, and for the audit log, audit logs saved to VAST DataBase. Both are off by default. The endpoint is the same as for VAST DataBase. | Access | What to ask for |
If you'll use several services, send one request that combines the items from each page. Several pages ask for the same endpoint, user and keys; for example, the Catalog uses the VAST DataBase endpoint, and reading Event Broker topics as tables uses the VAST DataBase setup.
Check that it works
Each setup or access page has a short check you can run once your platform team is done: VAST DataBase, Vector Search, Event Broker, DataEngine, DataStore and Catalog. If a check fails, the same page says what to look at, or links the section's errors or troubleshooting page.
See also
- Choose a service and choose a client.
- Automation tools: if your platform team provisions these items from code, such as Terraform or
vastpy.