Skip to main content
PySparkSnowflakeDatabricks

Mounting Cloud Storage — ADLS, S3 & GCS in Databricks

Mounting Cloud Storage — ADLS / S3 / GCS​

🌟 Simple Explanation (For Everyone)​

Imagine Databricks is your computer, and your cloud storage (ADLS, S3, GCS) is like an external hard drive.

To use the files easily, you “plug in” the storage.
In Databricks, this “plugging in” is called mounting.

After mounting, instead of using long, complicated cloud URLs, you get a simple folder:


/mnt/raw

This makes your life easier when reading or writing files.

🤔 Why Do We Mount Storage?​

Because mounting:

  • Makes paths shorter and easy to remember
  • Helps teams work in a clean, organized way
  • Lets you use cloud storage like a normal folder
  • Avoids repeating long URLs in every notebook

Think of it like giving your cloud data a nickname.

🟦 Mounting Azure ADLS (Easy Version)​

What you need:​

  • A service principal (like a username + password for apps)
  • Secrets stored safely in Databricks

Example (simple):​

# Connect details for ADLS
configs = {
  "client-id": "<client-id>",
  "client-secret": dbutils.secrets.get("storage", "adls-secret"),
  "tenant-id": "<tenant-id>"
}

## Mount ADLS to /mnt/raw
dbutils.fs.mount(
  source = "abfss://raw@yourstorage.dfs.core.windows.net/",
  mount_point = "/mnt/raw",
  extra_configs = configs
)

Now you can access files like:

/mnt/raw/customers.csv

Easy!

🟧 Mounting AWS S3 (Easy Version)​

For beginners, the easiest method:​

dbutils.fs.mount(
  source = "s3a://your-bucket-name",
  mount_point = "/mnt/mybucket",
  extra_configs = {
    "aws-access-key": dbutils.secrets.get("aws", "access-key"),
    "aws-secret-key": dbutils.secrets.get("aws", "secret-key")
  }
)

After this, you can access:

/mnt/mybucket/orders.json

🟥 Mounting Google Cloud Storage (Easy Version)​

You need:​

  • A service account JSON file (stored safely in secrets)
dbutils.fs.mount(
  source = "gs://your-gcs-bucket",
  mount_point = "/mnt/gcs",
  extra_configs = {
    "gcs-service-json": dbutils.secrets.get("gcp", "service-json")
  }
)

🧹 How to Unmount​

If you no longer want the mount:

dbutils.fs.unmount("/mnt/raw")

👍 Best Practices (Simple & Clear)​

✔ Don’t store passwords in notebooks​

Always use secrets.

✔ Use clean folder names​

Examples:

/mnt/raw
/mnt/bronze
/mnt/silver
/mnt/gold

✔ Mount storage for shared data​

Not for personal workspaces.

✔ Use direct cloud paths if mount is not needed​

Mounting is helpful—but not required.

🎉 Summary (In One Minute)​

  • Mounting = making cloud storage appear like a normal folder
  • It makes working with data much easier
  • You can mount ADLS, S3, and GCS
  • Use secrets to stay secure
  • Use simple, clean folder structures

Mounts are the first building block of working smoothly in Databricks. Once this is done, you’re ready for ingestion, Delta Lake, and the Medallion Architecture.

🚀 Coming Next​

👉 Autoloader — CloudFiles Ingestion End to End