Skip to content

docs: add Azure Files support as persistent storage option #55862

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

docs: add Azure Files support as persistent storage option #55862

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 20 additions & 0 deletions doc/source/train/user-guides/persistent-storage.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,6 +82,26 @@ Ensure that all nodes in the Ray cluster have access to the shared filesystem, e
so that outputs can be saved to there.
In this example, all files are saved to ``/mnt/cluster_storage/experiment_name`` for further processing.

Azure Files
~~~~~~~~~~~

Ray supports **Azure Files** as a persistent storage option for checkpoints, logs, and artifacts.

**Setup**

1. Create an Azure Storage Account and a File Share.
2. Mount the file share on each Ray node (e.g., ``/mnt/azurefiles``).
3. Configure Ray to write to this location by setting:

.. code-block:: bash

export RAY_STORAGE_PATH=/mnt/azurefiles
Comment on lines +94 to +98
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Aug 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The documentation suggests using export RAY_STORAGE_PATH=/mnt/azurefiles to configure the storage path for Azure Files. However, RAY_STORAGE_PATH does not appear to be a valid environment variable used by Ray Train for this purpose.

Using RunConfig(storage_path=...) is the correct and consistent way to configure a shared filesystem path, as shown in other parts of this document (e.g., for NFS/HDFS). Please update the configuration instructions to use RunConfig.

Suggested change
3. Configure Ray to write to this location by setting:
.. code-block:: bash
export RAY_STORAGE_PATH=/mnt/azurefiles
3. Configure your Ray Train or Tune script to use this path via ``RunConfig(storage_path)``:
.. code-block:: python
# In your Trainer/Tuner setup
run_config=train.RunConfig(storage_path="/mnt/azurefiles")


**Notes**

- Ensure the share is mounted with read/write permissions on all nodes (head + workers).
- For Kubernetes deployments, mount the Azure Files share as a volume on every pod and set ``RAY_STORAGE_PATH`` as an environment variable.
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Aug 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

Similar to the main setup instructions, the note for Kubernetes deployments incorrectly references the RAY_STORAGE_PATH environment variable. The correct approach is to configure the mounted volume path using RunConfig(storage_path=...) in the training script.

Suggested change
- For Kubernetes deployments, mount the Azure Files share as a volume on every pod and set ``RAY_STORAGE_PATH`` as an environment variable.
- For Kubernetes deployments, mount the Azure Files share as a volume on every pod and configure ``RunConfig(storage_path=...)`` to point to the mount path.

- For high-throughput training, consider using Premium tiers of Azure Files.

Local storage
-------------
Expand Down