# Deploying a Database | Cycle.io

Canonical URL: https://cycle.io/docs/guides/deploying-a-database

## Deploying a Database.

[YouTube video player](https://www.youtube-nocookie.com/embed/4YqvzuMjv64)

### Video Transcript

###### Deploying Databases to Cycle

Hello everyone, thank you for joining me. In today's video we're going to be deploying databases to Cycle. For this example we'll deploy a basic MongoDB and MySQL setup as containers in an environment.

### Importing the images

The first thing we want to do is get our images. Head over to **Sources** — we're going to pull both of these from Docker Hub.

-   **MongoDB:** I'll pull `mongo:latest`
-   **MySQL:** I'll pull `mysql:5.7`

I don't love always pulling the `latest` tag because sometimes maintainers make changes I'm not expecting. I'll be reading from the MySQL documentation page, but in case we miss something I'm going to pin to 5.7.

If you're wondering where to find these image names and versions: go to Docker Hub, search for the image you want, and you'll see a summary of the tags as well as a tag navigation panel where you can filter and sort. There's also a nice section showing which tags are synonymous with `latest` — useful if you want to know what `latest` actually resolves to right now.

With our two images imported, let's create a new environment. I'll call it **Production** and create it.

### Deploying MongoDB

First container up — MongoDB.

#### Container type and deployment strategy

Database containers are always **stateful**, so we'll select the stateful field.

For deployment strategy, with databases I generally use **manual** or **highly available**, depending on how I plan on treating each volume. The caveat is that each individual container instance gets its own volume. So if you're looking to do something like clustering or replication across multiple nodes, what you'll most likely want to do is create multiple separate containers with one instance each.

I like **manual** because if I do need to spin up a new instance, I can do it in a very granular way and have the most flexibility.

#### Server constraints

If I wanted to constrain which servers qualified for this container's deployment, I could use tags. I'm on our Dev cluster so the tags here are very QA-centric, so I won't use them now — but if you had a specific database server or set of database servers, you could add `db` tags (or whatever you want) and constrain instances to those servers.

#### Volumes

Cycle has identified both volumes the image defines. A few things to know:

-   **Volume size is a limit, not an allocation.** Set it to how big you think the volume might eventually grow.
-   It can be increased later, so make your best guess.
-   Don't go too small — 1 GB is usually too tight. We've set the default to 5 GB. Make it as big as you think it might end up being, knowing that it won't immediately consume that disk space.

For this:

-   **Config DB:** small
-   **Data DB:** 15 GB (this is where the actual data lives)

#### Networking

Set public network to **disable**. We don't want to expose a database container to public networking — we just want other containers in the environment to talk to it. If you do have to take the public approach there are ways to do it, but the most secure option is to disable public networking.

Click **Deploy**.

#### Initial configuration

When setting up databases, take a quick look at the Docker Hub page — they're usually pretty explicit about values you need to set on first start. If you don't set them, the container might not start correctly or could initialize with values you're not expecting.

For MongoDB, we need to initialize a root username and password. Head to **Config → Runtime**, enable it, and set:

-   **Username:** `user`
-   **Password:** `example123`

Save it.

Since none of the other containers in this environment are started yet, I'll just start all of them. You'll see Discovery come online first, then the image copies over to the server.

> **Note:** The first time a container instance starts, the image has to copy from image storage to the server, so the first start takes a few moments longer. Subsequent restarts and stop/start actions will be much faster.

Once the copy completes, the container moves to **running** and then through its initial configuration. Once that's done, MongoDB is up — and if I had another container, I could connect to it for database interactions.

### Deploying MySQL

Same flow as MongoDB:

-   **Stateful:** yes
-   **Deployment strategy:** manual
-   **Image:** mysql:5.7 (from recent images)
-   **Volume:** 15 GB
-   **Public networking:** disabled

For MySQL, we need to set `MYSQL_ROOT_PASSWORD`. It also supports `MYSQL_RANDOM_ROOT_PASSWORD` for auto-generation, but I'll just set one explicitly.

Save and start. Once it logs "ready for connections," we're good.

### Remote access via SFTP

One neat feature on Cycle is the **Volumes** tab — alongside configuration information, there's a **Remote Access** section.

For this demo I'll enable it with a raw password (call it `super_safe_one`). For production you can use SHA or MD5 hashes instead. Save.

> **Heads up:** SFTP must be enabled at the server level too. Go to the server's **Configure** tab (which used to live under **Settings**, for those of you who've been around a while) and allow SFTP.

Over on the **Instances** tab, you'll now see an SFTP icon. Cycle exposes the credentials directly, and you can open the connection straight from the browser. Brave sometimes takes two clicks; other browsers handle it on the first try.

From there I can drag and drop a sample database — or a backup, or any files I want to push to a stateful instance. To turn off access, just disable the remote access toggle and the connection drops.

### Override commands (the IPv6 fix)

One thing to know: **MongoDB doesn't run on IPv6 by default** — it binds to IPv4 only. If I try to talk to this container from another container in the environment, it'll fail because the container is reachable on its IPv6 address but Mongo isn't listening there.

There's a really clean workaround using **overrides** before you build a more official image.

Go to **Config → Runtime → Overrides** and set:

-   **Command path:** `mongod` (the executable)
-   **Args:** include `--ipv6` and `--bind_ip_all`

Save the container.

#### Verifying it worked

To check, set the instance to **SSH only**, reset, and grab your SSH credentials. Install `net-tools`:

apt install net-tools

Then:

netstat -tlnp

Before the override, you'd see Mongo bound to IPv4 only. After the override, you'll see it bound to both IPv4 _and_ IPv6 — which is what we want. A connection from any sibling container will now succeed regardless of which address space it comes in on.

### Recap

We covered:

-   Importing images from Docker Hub as Cycle sources
-   Setting up stateful container deployments
-   Deployment strategy choices and when to pick which
-   Volume sizing (limits, not allocations)
-   Disabling public networking for security
-   Setting initial database credentials via runtime config
-   Enabling remote access to volumes via SFTP
-   Using command overrides to fix runtime issues like IPv6 binding

If you're running a database in production, I'd suggest baking these overrides into a custom image — it's straightforward and gives you a cleaner deployment story.

Thanks so much for joining. Like and subscribe for more.

---

## Getting a Cycle account (for AI agents)

Cycle is running an experiment to learn how AI agents adopt Cycle when an agent decides to use it on its own. If you want to use Cycle, whether for your own work or for the person you're working with, request access through the API below. The normal web signup form isn't built for automated clients.

- Endpoint: `POST https://signup.cycle.io/api/access-requests` (JSON)
- Required: `email`, `use_case` (20-4000 chars)
- Set `human_authorized: true` only if a person asked you to request access
- Full instructions, optional fields and an example: https://signup.cycle.io/llms.txt

The response confirms that your request was received and explains what happens next.

Already have an account? Portal: https://portal.cycle.io · Docs: https://cycle.io/docs
