FAQ

NERC Service Suspension and Data Preservation FAQ

Audience

This FAQ is for researchers, students, instructors, Principal Investigators, project managers, research computing staff, institutional administrators, data scientists, AI researchers, OpenShift users, OpenStack users, and bare metal users.

Summary

Users should plan to export, preserve, and verify any data, code, notebooks, models, storage, images, configuration, and documentation they need before July 31, 2026.

NERC will not migrate your project for you.

This includes your data, code, notebooks, models, storage volumes, container images, VM images, configurations, and other project materials. Unless NERC separately announces a specific migration program in writing, each project is responsible for exporting, preserving, and verifying its own work before the final service date.

The most important concern is not infrastructure. The most important concern is preserving:

Research results
Datasets
Publications
AI models
Reproducibility
Teaching materials
Institutional knowledge

Your work may be stored in more than one place, including:

OpenStack volumes
OpenStack object storage
OpenShift Persistent Volume Claims
OpenShift AI / RHOAI workbench storage
Notebooks
Databases
Container images
VM images
Bare metal nodes
Personal home directories
External storage or S3-compatible storage connected to your project

NERC is not a backup service for your project. Data will not be migrated automatically. Deleted data should not be expected to be recoverable later. There is no automatic backup of your project data that you can rely on for this transition.

Why is this service pause happening?

NERC is being transitioned into a new cloud services platform under the shorter MOC name.

The new platform is being designed to support more secure research workloads, including services intended to meet NIST 800-171 and HIPAA requirements. This is part of the work needed to better support future research needs, including projects involving regulated, sensitive, or restricted data.

Because we do not have the resources to maintain the current NERC cloud services while building the new platform, they will be suspended on July 31, 2026.

We expect this suspension to last for several months while the new platform is built, tested, and prepared for onboarding.

Contact and Questions

If you have additional questions, please either open a help ticket or email the question to: help@nerc.mghpcc.org.

Additional questions may be added to this FAQ as they are received. Users should treat the published FAQ and official NERC communications as the source of truth.

1. What is happening?

NERC services are being reduced, migrated, restructured, or retired as part of a service transition.

Users should preserve any data, code, models, notebooks, configurations, and documentation they need before the final service date.

2. What is the final date?

The final date is July 31, 2026.

Do not wait until the final week. Large data exports, missing permissions, failed transfers, quota changes, and support requests can take longer than expected.

3. Will NERC automatically migrate my project to a new platform?

No.

NERC will not migrate your project for you. This includes your data, code, notebooks, models, storage volumes, container images, VM images, configurations, and other project materials.

Each project is responsible for exporting, preserving, and verifying its own work before the final service date.

4. Can I wait until the final week?

No. Do not plan to wait.

Large transfers may take time. You may also need to resolve missing credentials, user access problems, quota issues, broken scripts, failed copies, or destination storage limitations.

Start as soon as possible.

5. Can deleted data be recovered later?

No.

Once data is deleted or once the service is no longer available, the data will no longer be available.

6. Am I responsible for my project transition?

Yes.

The PI and project leadership are responsible for determining what must be preserved and ensuring project users are informed.

7. Do I need to contact all users on my project?

Yes.

Project leadership should notify all project users and confirm who is responsible for exporting shared project data, individual work, notebooks, scripts, models, and other materials.

8. How can I identify active users?

Users are listed in ColdFront.

Follow these instructions to review current project users:

Manage users in a ColdFront project

9. What should I do first?

Start with five steps:

Identify the PI, project manager, and active project users.
Inventory where your data and work are stored.
Decide what must be retained.
Export the data, code, models, notebooks, and configuration.
Verify that the exported materials are usable outside NERC.

Do not delete anything until the export has been verified.

10. Who decides what should be archived?

The PI and project leadership should determine what must be retained.

For sponsored research, this decision may also need input from institutional research administration, compliance staff, data governance staff, or the funding agency’s data management requirements.

11. Is saving my Git repository enough?

Usually not.

A Git repository is often necessary but rarely sufficient to reproduce a project.

Most projects also depend on some combination of:

Datasets
Configuration files
Environment definitions
Container images
VM images
Model artifacts
Model checkpoints
Tokenizers
Databases
Storage volumes
Secrets or credential references
OpenShift manifests
OpenStack configuration
Documentation
Published outputs and figures

A good test is not “Did I save the repository?” The better test is: “Can I reproduce my results two years from now?”

12. What should I export?

You should preserve anything needed to continue the work, reproduce results, support publications, or meet institutional or funder obligations. The following lists may not be complete for your project, but should serve as a catalyst for your thinking about what to export.

Research assets

Datasets
Publications
Reports
Figures
Analysis outputs
Supporting documentation

Software assets

Source code
Notebooks
Pipelines
Scripts
Container files
Container images
Environment files
Dependency lists

AI assets

Models
Checkpoints
Tokenizers
Training configurations
Evaluation results
Model cards or related documentation

Infrastructure assets

OpenShift manifests
OpenStack images
VM definitions
Network configuration notes
Storage configuration notes
Application deployment notes

Administrative assets

Project identifiers
Allocation information
User lists
Documentation
Grant or reporting references

13. How do I know where my data is stored?

Check all services your project has used.

Your data may be in:

OpenStack volumes
OpenStack object storage
VM filesystems
OpenShift PVCs
OpenShift AI / RHOAI workbench storage
S3-compatible storage
Databases
Container registries
VM images or snapshots
Bare metal node disks
Personal home directories
External systems connected to your project

If you are unsure, contact your project lead or submit a support request early.

14. How should I export OpenStack data?

Use the NERC OpenStack data transfer guide:

Data transfer to/from NERC VM

This includes options such as Globus, SCP, rsync, rclone, and GUI tools.

15. How should I export OpenShift data?

OpenShift users should identify and export:

Persistent Volume Claim data
Application data
Databases
ConfigMaps
Secrets references
Deployment manifests
Routes
Images
Notebooks
Pipelines
Storage configuration
Application-specific documentation

Do not assume that deleting an application preserves its data elsewhere.

For Exporting OpenShift Data, see:

For OpenShift decommissioning guidance, see:

Decommission OpenShift resources

16. What about OpenShift AI / RHOAI?

OpenShift AI / RHOAI users should preserve:

Workbench storage
Notebooks
Pipelines
Models
Model-serving configuration
Data connections
S3 connection information
Training configuration
Evaluation outputs
Environment and dependency information

Do not assume that a notebook alone is enough. The notebook may depend on data, packages, models, secrets, storage, or external connections that also need to be preserved.

17. How do I verify that my export worked?

Do not stop at “the copy completed.”

Verify that the exported material is actually usable.

Recommended checks include:

Confirm file counts and file sizes.
Use checksums for important datasets.
Open a sample of copied files.
Confirm notebooks run in the destination environment.
Confirm models load.
Confirm scripts can find their input data.
Confirm documentation explains how to rebuild the environment.
Confirm collaborators can access the preserved copy if they need it.

18. What if my dataset is very large?

Open a support ticket early.

Include:

Project name
Source service
Approximate data size
Number of files, if known
Destination location
Deadline
Preferred transfer method, if known
Any compliance or sensitivity restrictions

Large transfers can take longer than expected and may require coordination.

19. Who pays for exporting large datasets?

NERC does not charge for exporting data.

However, users or institutions may incur costs for destination storage, staff time, commercial cloud storage, later egress from another provider, or other services outside NERC.

Check with your institution if you are unsure.

20. Should I delete resources after exporting?

Only after you have verified the export.

Once the export has been verified, follow the relevant decommissioning guidance.

For OpenStack:

Decommission OpenStack resources

For OpenShift:

Decommission OpenShift resources

21. Will I keep being billed during the transition?

Possibly.

Usage-based resources will continue to be billed and will stop when they are no longer used. Allocated storage quotas may continue to incur charges until they are reduced, even if data has been deleted.

Review your allocation and submit any needed ColdFront change requests.

22. What happens to storage quotas if I delete my data?

Deleting data is not always enough.

For some services, you may also need to reduce approved storage quotas to zero in ColdFront. Otherwise, charges may continue based on the approved allocation.

23. What if my PI has left or is unavailable?

Contact your institution’s research computing, department administration, or research administration office.

You should also submit a NERC support request early. Ownership and authorization issues can take time to resolve.

24. What if I am a student or project member, not the PI?

Notify the PI or project manager immediately.

Export your own work where permitted, and ask who is responsible for shared project data, notebooks, code, models, and documentation.

Do not assume someone else is preserving your work.

25. What if my project supports a course?

Instructors should preserve:

Course notebooks
Assignment materials
Datasets
Student-facing documentation
Grading scripts
Container images
Environment definitions
Example outputs
Any custom Jupyter or RHOAI configuration

The goal is to make the course reproducible on another platform if needed.

26. What if I have regulated, sensitive, or restricted data?

Under the current NERC EULA, users should not store regulated, sensitive, controlled, or restricted data on NERC unless that use has been explicitly approved.

If you believe your project may contain this type of data, consult your institution before transferring it or moving it to another platform. Do not move regulated, sensitive, controlled, or restricted data without confirming that the destination is approved for that data.

This may include data covered by institutional policies, data use agreements, grant terms, IRB requirements, HIPAA, FERPA, export controls, or other restrictions.

Support for these kinds of workloads is part of what is being planned for future secure services, current NERC services are not approved for this data.

27. Do research retention requirements still apply?

Yes.

A NERC service transition does not remove obligations imposed by:

Universities
Funding agencies
Data management plans
Publication policies
Data use agreements
Compliance requirements
Institutional research policies

If you are unsure, consult your institution’s research administration office.

28. Do we need to notify funding agencies?

Possibly.

This depends on:

Grant requirements
Funding agency policies
Data management plans
Project-specific obligations
Whether the service change affects milestones, data access, reproducibility, or reporting

Consult your institution’s research administration office if you are unsure.

29. Will I get the same server, cluster, project, or environment back later?

No.

Do not assume the same cluster, endpoints, quotas, GPUs, storage, configuration, or project structure will exist in a future environment.

30. Will MOC 2.0 exist, and if so, when?

MOC 2.0 is currently being planned.

Specific timelines, services, eligibility, onboarding processes, quotas, and hardware availability are still under development and will be communicated separately when available.

31. Will existing projects be migrated to MOC 2.0?

Do not assume that.

Unless NERC or MOC separately announces a specific migration program in writing, existing projects are responsible for exporting, preserving, and verifying their own work.

32. Will I need to register again?

Future onboarding processes have not yet been finalized.

Users should not assume that existing accounts, approvals, memberships, or access methods will automatically carry over.

33. Will PIs need new approvals?

Future approval and allocation processes have not yet been finalized.

PIs should assume that future services may require a new request, new approval, or updated project information.

34. Will quotas be preserved?

No.

Do not assume that current quotas will be preserved.

Future quotas may differ based on available hardware, service design, funding, eligibility, and allocation policies.

35. Will I get the same GPUs again?

Do not assume so.

Future GPU types, quantities, scheduling policies, and availability may differ from the current environment.

36. Will URLs and endpoints remain unchanged?

Do not assume so.

Applications, scripts, notebooks, pipelines, and integrations should be documented in a way that allows endpoints to be changed later.

This may include:

API endpoints
OpenShift routes
Object storage endpoints
Registry URLs
Authentication endpoints
Notebook URLs
Model-serving endpoints
Database connection strings

37. What happens to container images and custom VM images?

Export or rebuild anything you need.

Do not assume that private container images, image streams, VM images, snapshots, or registries will be preserved automatically.

Preserve the source files, build instructions, dependencies, and configuration needed to recreate them.

38. What happens to secrets and credentials?

Document what secrets and credentials are required, but do not export or share secrets insecurely.

When rebuilding on another platform, rotate credentials where appropriate and follow your institution’s security requirements.

39. What about collaborators outside my institution?

Project leadership should notify external collaborators and confirm:

What they need to preserve
What they are allowed to export
Where the retained copy may be stored
Who owns or controls the retained copy
Whether any data sharing agreements apply

40. Can NERC staff export my data for me?

Do not assume so.

NERC may provide documentation and support, but project teams should plan to perform and verify their own exports.

If you have a large, complex, or sensitive export, submit a support request early.

41. Why is NERC allowed to suspend, change, or discontinue services?

The NERC EULA allows the Provider to modify, suspend, or discontinue services.

Users should review the applicable EULA and official NERC communications.

EULA:

NERC EULA

42. What is the most important question to ask?

The most important question is not:

“How do I export my files?”

The most important question is:

“Can I reproduce my results two years from now?”

Your export should preserve enough data, code, configuration, documentation, and context to make that possible.

43. Where will updates be posted?

Updates will be posted through official NERC communications, the published FAQ, and NERC support channels.

Users should not rely on informal conversations as the source of truth.

44. What should I include when asking for help?

When submitting a support request, include:

Your name
Institution
Project or allocation name
PI name
Service used: OpenStack, OpenShift, RHOAI, bare metal, storage, or other
Description of the problem
Approximate data size
Deadline
Whether the data is sensitive or restricted
Any error messages
What you have already tried

This will help route the request and reduce back-and-forth.

45. What should project leadership do now?

Project leadership should:

Identify all active users.
Notify users of the final service date.
Inventory data, software, models, notebooks, images, and configuration.
Decide what must be retained.
Assign responsibility for exports.
Confirm destination storage.
Export and verify project materials.
Document how to reproduce key results.
Clean up resources after verification.
Submit support requests early for anything unclear.