# Data Transfer Guide: EOS ↔ PFS

This guide explains how to transfer data from **EOS** to **PFS** or from **PFS** to **EOS** using Argo Workflows.

You can use these workflows to schedule or automate periodic data movement jobs between your object storage buckets and parallel filesystem volumes.


## Prerequisites

Before running any workflow:

- You must have an existing **EOS** bucket. To create an EOS, follow this [link](/docs/tir/Datasets/QuickStart/#create-a-new-dataset)

- You must have a **PFS** filesystem created and mounted to an instance. To create a **PFS**, follow this [link](/docs/tir/pfs/)

- Ensure your access credentials:

    - `bucket_access_key`

    - `bucket_secret_key`

:::info
You can view the Bucket Name in the Dataset’s Overview tab.
:::


## Transfer Data: PFS → EOS

This workflow uploads files from your PFS to your EOS bucket.

### Required Parameters

| Parameter                 | Description                          |
| ------------------------- | ------------------------------------ |
| `source_fs_name`          | Name of the filesystem               |
| `source_path`             | Path inside the filesystem to upload |
| `destination_bucket_name` | EOS bucket name                      |
| `destination_path`        | Destination folder inside the bucket |
| `bucket_endpoint_url`     | EOS endpoint (default included)      |
| `bucket_access_key`       | Your EOS access key                  |
| `bucket_secret_key`       | Your EOS secret key                  |


### Create Pipeline

<a href="/yaml/fs-to-eos.yaml" download="fs-to-eos.yaml">
   <strong>Download fs-to-eos.yaml</strong>
</a>

Now create a **pipeline** using the `fs-to-eos.yaml` file. For more details about creating **pipeline** follow this [link](/docs/tir/Pipeline/QuickStart/#create-your-first-pipeline)

:::note
Enter the correct parameters in the YAML file, such as `source_fs_name`, `source_path`, and others. You may also fill them in manually while creating a **run** for this **pipeline**.
:::


Now create a **run** for the pipeline. While creating the run, make sure to provide the correct **parameters**; otherwise, they will be automatically fetched from the YAML file if already specified.

For more details about creating a **run** for the pipeline, follow this [link](/docs/tir/Pipeline/QuickStart/#create-a-run)

Once the run completes **successfully**, you can view the execution logs by clicking the run name.

## Transfer Data: EOS → PFS

This workflow downloads data from your EOS bucket into your PFS filesystem.

### Required Parameters

| Parameter             | Description                            |
| --------------------- | -------------------------------------- |
| `source_bucket_name`  | EOS bucket name                        |
| `source_path`         | Folder/object prefix in the bucket     |
| `destination_fs_name` | Target PFS filesystem name             |
| `destination_path`    | Folder inside filesystem to store data |
| `bucket_endpoint_url` | EOS endpoint                           |
| `bucket_access_key`   | EOS access key                         |
| `bucket_secret_key`   | EOS secret key                         |


### Create Pipeline

<a href="/yaml/eos-to-fs.yaml" download="eos-to-fs.yaml">
   <strong>Download eos-to-fs.yaml</strong>
</a>

Now create a **pipeline** using the `eos-to-fs.yaml` file. For more details about creating **pipeline**, follow this [link](/docs/tir/Pipeline/QuickStart/#create-your-first-pipeline)

:::note
Enter the correct parameters in the YAML file, such as `source_bucket_name`, `source_path`, and others. You may also fill them in manually while creating a **run** for this **pipeline**.
:::



Now create a **run** for the pipeline. While creating the run, make sure to provide the correct **parameters**; otherwise, they will be automatically fetched from the YAML file if already specified.

For more details about creating a **run** for the pipeline, follow this [link](/docs/tir/Pipeline/QuickStart/#create-a-run)

Once the run completes **successfully**, you can view the execution logs by clicking the run name.




---