> ## Documentation Index
> Fetch the complete documentation index at: https://docs.regatta.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# RegattaDB Lifecycle

> Commands for creating, starting, stopping, resetting, and cleaning a RegattaDB deployment.

These commands run the sequence of steps described in
[System Setup Overview](./overview.mdx). Each one takes `--name <config>` to select which
[config](../configuration/overview.mdx) to act on.

## Deploying RegattaDB

### One-step

```bash theme={null}
rdt create-cluster --name <config> --src <path/to/rpms> [--clean-install]
```

`--src` accepts either a `.tar.zst` archive or a directory containing Regatta RPMs.
`--clean-install` removes any previous installation on the nodes before reinstalling.

### Step-by-step

```bash theme={null}
rdt distribute --name <config> --src <path/to/rpms>
rdt setup --name <config> [--clean-install]
rdt start-services --name <config>
rdt configure --name <config>
rdt start-cluster --name <config>
```

This runs the same sequence as `create-cluster`, one step at a time. Useful if you want
to inspect or troubleshoot a specific stage.

A few things worth knowing:

* Management commands take effect on the System Manager as soon as they're issued.
  There's no transactional rollback. If `configure` fails partway through (for example,
  a node can't be reached), whatever was already applied before the failure stays applied.
* After `start-cluster`, if some modules didn't reach an active state, RDT prints them as
  a warning rather than failing outright. Check the warning and investigate those
  modules before considering RegattaDB ready.

## Starting and stopping

There are two independent layers you can start or stop:

* **`start-services`/`stop-services`**: starts or stops the System Manager and SNA as
  node-level services. This doesn't affect whether the database modules themselves are
  running.
* `start-cluster`**/**`stop-cluster`: tells the System Manager to start or stop the operational modules (RDB, GDD, DCM, Sequencer), bringing the database up or down. Under the hood these issue the `MANAGE SYSTEM START` and `MANAGE SYSTEM SHUTDOWN` management commands (see the [System States and Commands](https://docs.regatta.dev/sql/system-management/system-states-and-commands)).

```bash theme={null}
rdt stop-cluster --name <config>
rdt start-cluster --name <config>
```

## Resetting RegattaDB

```bash theme={null}
rdt reset --name <config>
```

Stops services, clears all RegattaDB data, and restarts services, but it does **not**
automatically reconfigure or restart the database afterward. Run `configure` and
`start-cluster` again once `reset` finishes:

```bash theme={null}
rdt reset --name <config>
rdt configure --name <config>
rdt start-cluster --name <config>
```

## Cleaning up

```bash theme={null}
rdt clean --name <config>
```

Removes Regatta's installation, files, logs, and traces from every node. If RDT itself
is installed on a given node, `clean` preserves RDT and its own logs on that node while
removing everything else; if RDT isn't installed there, the node is wiped completely.

## Collecting debug information

```bash theme={null}
rdt collect --name <config> --destination <path> [--nodes <node>...] [--remote-destination <path>]
```

Collects logs and diagnostics from RegattaDB nodes into a local directory. Use `--nodes`
(repeatable, by node name or id) to limit collection to specific nodes; omit it to
collect from all nodes. `--remote-destination` controls the staging directory used on
each node (defaults to `/tmp/regatta-collect`). Run `rdt collect --help` for the full
list of options for this command.

This command wraps the standalone `regatta-collect` tool that runs on each node. See
that tool's own documentation for its full set of options and additional examples.

## Next steps

See [Examples](../examples/quickstart-single-node.mdx) for end-to-end walkthroughs.