Importing and Exporting Data#

Role: Data Engineer · Researcher / Data Analyst (egress download only)

Task

Data Engineer

Researcher / Data Analyst

Import data via SFTP

Import data via Globus

Import data via allowlisted web portal

Stage files for egress review

Review and approve egress

Download approved egress files

All data movement into and out of the SDE must comply with Northwestern and NIST SP 800-171 standards and the DUA governing your project. Only Data Engineers can import data into the SDE. Researchers and Data Analysts who need data imported should contact their project Data Engineer.

Data can only leave the SDE after a Data Engineer reviews it for compliance and logs the egress. Researchers and Data Analysts can stage files for review but cannot download data directly from within the SDE.

Warning

Do not attempt to transfer data using methods not described on this page. Unauthorized transfer methods violate the SDE Acceptable Use Policy and may breach your project DUA.


Data Ingress#

Only data consistent with the DUA governing your project may be imported into the SDE. All ingress methods require IT configuration before they can be used. Ingress options are discussed during SDE onboarding. If you need an additional method or have a new ingress request, submit a support request so that it can be reviewed and configured.

The SDE supports three approved ingress methods:

Method

Best for

Requires IT setup

SFTP

Large or recurring transfers from a data provider’s server

Yes — firewall rules and key exchange

Globus

Transfers from Northwestern systems or other Globus endpoints

Yes — private collection provisioning

Allowlisted web portal

One-time downloads from a web interface (SharePoint, Dropbox, etc.)

Yes — URL and IP allowlisting

If your data cannot be transferred via any of these methods, contact Northwestern IT to discuss alternatives before attempting any transfer.


SFTP Transfer#

Role: Data Engineer

SFTP (Secure File Transfer Protocol) allows your data provider to push files directly to a VM in the Ingress Project over an encrypted connection. Because SDE VMs have no general internet access, Northwestern IT must open a specific firewall port and allowlist your data provider’s IP address before any transfer can occur. This is required by the NIST SP 800-171 network controls governing the SDE.

Note

SFTP is best suited for large datasets, recurring transfers, or providers who cannot use Globus or a web portal. The setup involves coordination between you, your data provider, and Northwestern IT.

What you need to collect from your data provider#

Before submitting a setup request to Northwestern IT, gather the following information from your data provider. You will need to include all of it in your request.

Information needed

What to ask your provider

Static IP address(es)

What is the static IP address or IP range your SFTP server will connect from?

SFTP port

What port does your SFTP server use?

Authentication method

Do you use SSH key authentication, password authentication, or both?

Public SSH key (if key-based)

Can you provide your public SSH key so we can authorize your connection?

Expected transfer schedule

Will transfers be one-time, recurring, or on-demand?

Approximate data volume

What is the expected size of each transfer?

Important

Your data provider must use a static IP address. Dynamic or shared IPs cannot be allowlisted. If your provider cannot provide a static IP, SFTP is not a viable option and you should consider Globus instead.

Key-based authentication#

The SDE uses SSH key-based authentication for SFTP connections. Password-based authentication is not permitted, as it does not meet NIST SP 800-171 access control requirements.

How this works:

  • Your data provider generates an SSH key pair (a private key they keep, and a public key they share with you).

  • Northwestern IT places the provider’s public key on the Ingress VM to authorize their connection.

  • When your provider connects, their private key is verified against the public key without any password being transmitted.

Ask your provider to send you their public key (typically a file ending in .pub). Include it in your IT setup request. Never ask for or accept their private key.

Submitting the setup request#

Once you have the information above, submit an ingress setup request to Northwestern IT. Include:

  • All information from the table above

  • Your data provider’s public SSH key (if using key-based auth)

  • The name of your SDE environment and project

Northwestern IT will configure the firewall rules, open the necessary port. You will need to generate the public key in the VM. You will be notified when the connection is ready to test.

Verify the connection#

Once IT confirms setup is complete:

  • Coordinate a test transfer with your data provider.

  • Ask them to push a small test file.

  • Verify it arrives in the expected location on the Ingress VM, then move it to the appropriate bucket in the Data Lake Project following your standard ingress workflow.

Common errors#

Error

Plain-English meaning

What to do

Connection refused

The firewall port is not open or the IP is not allowlisted

Confirm the provider is connecting from the correct IP and contact IT to verify firewall rules

Permission denied (public key)

The public key is not installed or does not match

Ask your provider to re-send their public key and submit an update request to IT

Connection timeout

The provider is hitting the wrong IP address or port

Confirm with IT the correct hostname and port to give your provider


Globus Transfer#

Globus is a secure, high-performance data transfer service used across research institutions. The SDE uses Globus to allow Data Engineers to bring data in from other Northwestern Globus collections or from collections at other institutions and data providers.

How the SDE Globus setup works#

Upon request, Northwestern IT provisions a private Globus collection scoped to your SDE environment. This collection:

  • Is accessible only to users within your specific SDE environment (authenticated via your Northwestern Google account).

  • Is isolated from other SDE environments — users in one enclave cannot access another enclave’s collection.

  • Deposits incoming data into a storage bucket in the Ingress Project.

Only Data Engineers can transfer data using this collection. Researcher/Data Analysts do not have access to the Ingress Project or the Globus collection.

Note

Globus is the preferred method for transferring data from other Northwestern systems, including Quest, RDSS, FSMResFiles, Northwestern OneDrive/SharePoint, and Amazon S3. See the Northwestern Globus documentation for the full list of available collections.

Requesting Globus setup#

Globus must be requested and configured before first use. Submit a Globus ingress request to Northwestern IT. Include:

  • The name of your SDE environment

  • The source collection you plan to transfer from (e.g., Quest, RDSS, an external institution’s collection)

  • Whether the transfer is one-time or recurring

IT will provision your private collection and notify you when it is ready.

Transferring data#

Once your collection is provisioned:

Note

Globus transfers are initiated and managed through the Globus web application, not the Google Cloud Console. The Console is used only to verify that data arrived in the Ingress Project bucket after the transfer completes.

Log in to Globus

  1. Go to app.globus.org .

  2. Select Northwestern University as your organization and sign in with your Northwestern credentials.

Find your SDE collection

  1. In the File Manager, search for your SDE environment’s private collection by name. Northwestern IT will provide the exact collection name when setup is complete.

  2. Your collection will only be visible to users within your SDE environment.

Set up the transfer

  1. In the left panel, navigate to your source collection (where the data is coming from).

  2. In the right panel, select your SDE private collection as the destination.

  3. Navigate to the correct destination folder within the Ingress bucket.

  4. Select the files or folders to transfer and click Start.

Verify the transfer

  1. Globus will send an email notification when the transfer completes.

  2. Log in to the Google Cloud Console, navigate to the Ingress Project, and confirm the files appear in the Ingress bucket under Cloud Storage > Buckets .

Verify the transfer#

  • Check the Globus notification email for any listed skipped files or warnings.

  • Open the transfer’s detail page in Globus Activity and confirm the file count and total size match, with no errors in the Events log.

  • Confirm the file count in the Ingress bucket via Cloud Storage > Buckets in the Ingress Project.

  • Spot-check a few file sizes against the source to rule out truncation or corruption.

  • Confirm no partial or .globus-* temp files remain.

  • Once verified: log the ingestion per your project’s intake process, move the data to the appropriate bucket in the Data Lake Project, and clean up temporary files from the Ingress bucket.

  • If any check fails, do not move the data forward — see Common errors or the Globus troubleshooting guide .

Common errors#

Error

Plain-English meaning

What to do

Collection not found

The SDE collection has not been provisioned yet or the name is wrong

Confirm with IT that setup is complete and get the exact collection name

Permission denied

Your account is not authorized for this collection

Contact IT to verify your account is linked to the correct SDE environment

Transfer failed or stalled

A network or endpoint issue interrupted the transfer

Globus will automatically retry; if it continues to fail, check the Globus troubleshooting guide


Allowlisted Web Portal#

If your data source provides a web-based download interface (such as SharePoint, Dropbox, REDCap, or a data repository portal), you can access it from a VM in the Ingress Project once the URL and IP addresses are allowlisted by Northwestern IT.

SDE VMs have no general internet access. IT must explicitly permit each external URL and its associated IP addresses before you can reach it from the Ingress VM. This ensures that only reviewed and approved external connections are permitted, consistent with NIST SP 800-171 network controls.

What you need to collect before requesting#

Before submitting a request, gather the following from your data source:

Information needed

How to find it

Full URL(s) of the portal

The exact web address(es) you need to access to download data

IP address(es) of the portal

Ask your data provider: “What are the IP addresses for [portal URL]?” Or use a tool like nslookup or dig on your local machine

Whether the portal redirects to additional domains

Browse the portal on your local machine and note any additional domains the download process involves (e.g., CDN or authentication redirects)

Important

Web portals frequently redirect through multiple domains during login and download flows. If you provide only the main URL and the download involves redirects to other domains, the transfer will fail mid-process. Browse the full download flow on your local machine first and note every domain involved. When in doubt, contact IT early — they can assist in identifying the full set of addresses that need to be allowlisted.

Submitting the allowlist request#

Submit an ingress setup request to Northwestern IT. Include:

  • The full URL(s) of the portal

  • The IP address(es) associated with the portal

  • Any additional domains involved in the login or download flow

  • The name of your SDE environment

IT will review the request and add the approved URLs and IP addresses to the allowlist. You will be notified when the configuration is complete.

Accessing the portal and downloading data#

Once IT confirms the allowlist is in place:

Note

This procedure uses the VM in the Ingress Project, accessed via SSH-in-browser or the VM desktop. You cannot access web portals from your local machine and then move the data in — the portal must be accessed from within the SDE.

Start the Ingress Project VM

  1. Log in to the Google Cloud Console and select the Ingress Project.

  2. Go to Compute Engine > VM Instances and start the VM.

Access the portal and download data

  1. Open SSH-in-browser on the Ingress VM.

  2. Open a web browser from the VM desktop (if using VM Desktop) and navigate to the allowlisted URL, or use a command-line download tool such as wget or curl from the SSH-in-browser terminal.

  3. Log in to the portal and download the files. Files will be saved to local VM storage.

Move data to the Ingress bucket

  1. Once the download is complete, transfer the files from local VM storage to the Ingress Project storage bucket:

gcloud storage cp /path/to/downloaded/file gs://INGRESS_BUCKET_NAME/

Placeholder

What to substitute

/path/to/downloaded/file

The path to the file on the VM (use pwd to find your current directory)

INGRESS_BUCKET_NAME

The name of the Ingress Project storage bucket

Expected output:

Copying file:///path/to/downloaded/file to gs://INGRESS_BUCKET_NAME/file
  Completed files 1/1

Verify

  1. Confirm the file appears in the bucket:

gcloud storage ls gs://INGRESS_BUCKET_NAME/

Verify the data#

  • Data entering the VM is scanned with Crowdstrike, log the ingestion, and move the data to the appropriate bucket in the Data Lake Project.

  • Remove the local copy from the VM and the staging copy from the Ingress bucket after the data has been moved.

Common errors#

Error

Plain-English meaning

What to do

Page does not load / connection timeout

The URL or IP is not yet allowlisted

Confirm IT has completed the configuration and that you are using the exact URL from your request

Login succeeds but download fails

The download redirects to a domain that is not allowlisted

Note the domain in the error or browser, and submit an updated request to IT to add the additional address

gcloud storage cp fails with AccessDeniedException

Your account does not have write access to the Ingress bucket

Contact IT to verify your IAM permissions on the Ingress Project


Data Egress#

Role: Data Engineer · Researcher / Data Analyst

Data can only leave the SDE after a Data Engineer reviews it for compliance with the project DUA and logs the egress. Files containing PII or other restricted information as defined by your DUA may not be exported.

Warning

Exporting data that has not been reviewed and approved by a Data Engineer is a policy violation and may breach your project DUA. Use only the procedures described on this page.

The SDE supports three approved egress methods:

Method

Best for

Requires IT setup

Egress bucket download

Downloading files to your managed endpoint via the Console

No — buckets are managed by the Data Engineer

SFTP

Pushing approved files to a recipient’s SFTP server

Yes — firewall rules and key exchange

Globus

Transferring approved files to a Northwestern or external Globus collection

Yes — separate egress collection provisioning

The egress workflow involves both roles and must follow this sequence in order:

Step

Who does it

Where

1. Copy file(s) to Egress Dataprep bucket in the Data Lake and notify Data Engineer

Researcher / Data Analyst or Data Engineer

VM (SSH-in-browser)

2. Review the file for compliance with the DUA

Data Engineer

Data Ops VM (SSH-in-browser)

3. Submit an IT ticket to log the egress event

Data Engineer

Service request

4. Copy the approved file from the Data Lake to the Egress Project bucket

Data Engineer

Data Ops VM (SSH-in-browser)

5. Notify the researcher the file is ready and provide the bucket name

Data Engineer

6. Download the file from the Egress Project bucket to your managed endpoint

Researcher / Data Analyst or Data Engineer

Google Cloud Console

7. Delete the data from the egress bucket within 3 days of download

Data Engineer

Google Cloud Console

Attention

Data must be moved from the Data Lake to the Egress Project bucket by the Data Engineer before a download is possible. Steps 1 through 5 must be complete before attempting to download.


Stage a file for egress review#

Role: Researcher / Data Analyst · Data Engineer

Move the file you want to export to the Egress Dataprep bucket in the Data Lake Project. This signals to the Data Engineer that the file is ready for compliance review.

Note

These commands run on the VM. Make sure your VM is running and you have a terminal open via SSH-in-browser before proceeding.

Copy the file to the Egress Dataprep bucket

gcloud storage cp gs://SOURCE_BUCKET/FILE_NAME gs://EGRESS_DATAPREP_BUCKET/

Placeholder

What to substitute

SOURCE_BUCKET

The bucket where your file currently lives

FILE_NAME

The name of the file you want to stage for egress

EGRESS_DATAPREP_BUCKET

The name of the Egress Dataprep bucket in the Data Lake Project (your Data Engineer can provide this name)

Expected output:

Copying gs://SOURCE_BUCKET/FILE_NAME to gs://EGRESS_DATAPREP_BUCKET/FILE_NAME
  Completed files 1/1

Verify

gcloud storage ls gs://EGRESS_DATAPREP_BUCKET/

Confirm your file appears in the listing.

Important

After moving the file, notify your project Data Engineer that a file is ready for egress review. Include the file name and a brief description of its contents.

The Data Engineer will review the file, submit a compliance ticket to IT, and then move the approved file to the Egress Project bucket. You will not be able to download the file until the Data Engineer completes those steps and notifies you that the file is ready.


Review and approve egress#

Role: Data Engineer

Files staged by Researchers or Data Analysts appear in the Egress Dataprep bucket in the Data Lake Project. Use the VM in the Data Ops Project to review each file for compliance with the project DUA.

Review checklist:

  • Does the file contain PII or other information restricted by the DUA?

  • Has aggregation or de-identification been applied correctly?

  • Is the egress request consistent with the project’s approved use?

Warning

Do not approve files that contain restricted information. If you are unsure whether a file is safe to export, consult your institution’s data governance contact before proceeding.

You must maintain a log of all approved egress requests. The log must include at minimum:

  • Date of egress

  • File name and description

  • Who requested the egress

  • Your review decision and rationale

Submit an IT ticket to log the egress event#

Once you have reviewed and approved the file, submit an egress notification ticket to Northwestern IT before moving the data to the Egress Project. This ticket does not require approval — it is a required compliance log of the egress event.

The ticket should include:

  • File name and a brief description of the contents

  • Requesting user’s name

  • Date of your review

  • The egress method being used (bucket download, SFTP, or Globus)

  • Confirmation that the file complies with the project DUA

Important

Do not proceed to move the file until the ticket has been submitted. The ticket is the official record that this egress occurred and may be required for compliance audits.


Place approved files in the egress bucket#

Role: Data Engineer

The Egress Project contains dedicated egress buckets. Buckets are persistent, and they do not need to be recreated for each request. Only the data inside them must be deleted within 3 days of download.

Each egress bucket is scoped to a specific researcher. Access must be granted individually and users should not add multiple researchers to the same bucket, as each researcher should only be able to see their own approved files.

Create or confirm the egress bucket

If this is the first egress request for this researcher, create a new bucket in the Egress Project. Name it clearly so it is identifiable by researcher or workspace, for example: test-egress-WORKSPACENAME.

If the researcher already has a bucket from a previous request, you can reuse it — confirm it is empty before adding new files.

Grant access to the specific researcher

Add the researcher to the bucket with Storage Object Viewer access. Use their Northwestern email address. Do not add any other users to this bucket.

Important

Access must be granted to the specific requesting researcher only. Each researcher should only have visibility into their own egress bucket. If multiple researchers in the same environment have egress requests, each must have their own separate bucket.

Once the bucket is ready and access is confirmed, copy the approved file into it.

Note

These commands run on the VM in the Data Ops Project. Make sure your VM is running and you have a terminal open via SSH-in-browser before proceeding.

Copy the approved file to the egress bucket

gcloud storage cp gs://EGRESS_DATAPREP_BUCKET/FILE_NAME gs://EGRESS_BUCKET/

Placeholder

What to substitute

EGRESS_DATAPREP_BUCKET

The Egress Dataprep bucket in the Data Lake Project

FILE_NAME

The approved file name

EGRESS_BUCKET

The researcher’s egress bucket in the Egress Project

Expected output:

Copying gs://EGRESS_DATAPREP_BUCKET/FILE_NAME to gs://EGRESS_BUCKET/FILE_NAME
  Completed files 1/1

Verify

gcloud storage ls gs://EGRESS_BUCKET/

Confirm the file appears before notifying the researcher.

Important

Notify the researcher that their file is ready to download and provide the bucket name. Remind them that the data must be deleted from the bucket within 3 days of download.


Download approved egress files#

Role: Researcher / Data Analyst · Data Engineer

Once a Data Engineer notifies you that your file is approved and ready, download it from the Egress Project bucket to your managed Endpoint.

Open the Egress Project bucket

  1. Log in to the Google Cloud Console.

  2. Switch to the Egress Project in the project dropdown.

Selecting the Egress Project in the Console
  1. Go to Cloud Storage > Buckets and click the bucket name provided by your Data Engineer.

Selecting the egress bucket

Download the files

  1. Browse the bucket to find your files.

  2. Check the box next to the file(s) you want to download.

  3. Click the Download button and save the files to your managed Endpoint.

Download button in Cloud Storage

Verify

  1. Confirm the files are intact on your local machine. Notify your Data Engineer that the download is complete.

Attention

The data in the egress bucket must be deleted within 3 days of download. Anyone with access to the bucket can delete the files.


SFTP Egress Transfer#

Role: Data Engineer

SFTP egress allows you to push approved files from the Data Ops VM directly to a recipient’s SFTP server over an encrypted connection. As with ingress, SDE VMs have no general internet access — Northwestern IT must open a specific firewall port and allowlist the recipient’s server IP before any transfer can occur.

Note

SFTP egress is best suited for recipients who cannot receive files via Globus or a direct download, or where your DUA requires server-to-server transfer. The setup requires coordination between you, the recipient, and Northwestern IT.

What you need to collect from your recipient#

Before submitting a setup request to Northwestern IT, gather the following from your data recipient:

Information needed

What to ask your recipient

SFTP server hostname or IP address

What is the hostname or static IP address of your SFTP server?

SFTP port

What port does your SFTP server use?

SFTP username

What username should we use to authenticate to your server?

Authentication method

Do you accept SSH key authentication or password authentication?

Their public SSH key or password setup

Can you create an account for our transfer and share the credentials securely, or shall we provide a public key for key-based auth?

Expected transfer schedule

Will this be a one-time transfer or recurring?

Important

The recipient’s SFTP server must have a static IP address. Dynamic IPs cannot be allowlisted. If the recipient cannot provide a static IP, discuss an alternative method with IT.

Key-based authentication#

SSH key-based authentication is the preferred method for SFTP egress. In this model:

  • Data Engineer generates an SSH key pair on the Data Ops VM.

  • The public key is shared with your recipient, who installs it on their SFTP server to authorize the connection.

  • The private key remains on the VM and is never shared. Ask your recipient to accept a public key and install it on their server. IT will provide the public key when they configure the connection. Never transmit the private key.

Submitting the setup request#

Submit an egress setup request to Northwestern IT. Include:

  • All information from the table above

  • The name of your SDE environment

  • The files or dataset you are planning to transfer (for compliance review)

  • Northwestern IT will configure the firewall rules, open the necessary port. You will be notified when the connection is ready to test.

Transferring files#

Once IT confirms setup is complete, files must already be in an egress bucket in the Egress Project and reviewed before transfer. Use the Data Ops VM to push the files to the recipient’s server.

Note

These commands run on the VM in the Data Ops Project. Make sure your VM is running and you have a terminal open via SSH-in-browser before proceeding.

Copy the file from the egress bucket to the VM

gcloud storage cp gs://EGRESS_BUCKET/FILE_NAME /tmp/FILE_NAME

Placeholder

What to substitute

EGRESS_BUCKET

The egress bucket in the Egress Project

FILE_NAME

The approved file name

Push the file to the recipient’s SFTP server

sftp -i ~/.ssh/PRIVATE_KEY_FILE SFTP_USERNAME@SFTP_HOST

Placeholder

What to substitute

PRIVATE_KEY_FILE

The private key file on the VM (IT will provide the filename)

SFTP_USERNAME

The username provided by your recipient

SFTP_HOST

The recipient’s SFTP server hostname or IP address

Once connected, at the sftp> prompt:

put /tmp/FILE_NAME DESTINATION_PATH/

Placeholder

What to substitute

FILE_NAME

The file you copied to /tmp in Step 1

DESTINATION_PATH

The destination directory on the recipient’s server

Type exit to close the SFTP session.

Verify and clean up

Confirm with your recipient that the file arrived intact. Then remove the local copy from the VM:

rm /tmp/FILE_NAME

Common errors

Error

Plain-English meaning

What to do

Connection refused

The firewall port is not open or the server IP is not allowlisted

Confirm the recipient is using the correct IP and contact IT to verify firewall rules

Permission denied (public key)

The public key is not installed on the recipient’s server or does not match

Ask your recipient to re-verify the key installation and contact IT if needed

Connection timeout

The wrong hostname or port is being used

Confirm the server details with your recipient and verify with IT


Globus Egress Transfer#

Role: Data Engineer

Globus egress allows you to push approved files from the SDE to a recipient’s Globus collection, whether at Northwestern or an external institution. The SDE uses a separate Globus collection for egress, distinct from the ingress collection. Both must be requested and provisioned independently.

Note

The egress Globus collection is scoped to your SDE environment and authenticated via your Northwestern Google account. Only Data Engineers can transfer data out using this collection.

Requesting egress Globus setup#

If you have not already requested an egress Globus collection, submit an egress setup request to Northwestern IT. Include:

  • The name of your SDE environment

  • The destination collection you plan to transfer to (e.g., a collaborator’s institutional Globus endpoint)

  • Whether the transfer is one-time or recurring IT will provision your egress collection and notify you when it is ready.

Transferring files#

Files must be in the egress bucket in the Egress Project and reviewed before initiating a Globus transfer. The Globus transfer is then initiated from the Globus web application.

Note

Globus transfers are initiated through the Globus web application, not the Google Cloud Console. The Console is used only to confirm files are in the egress bucket before starting the transfer.

Create or confirm the egress bucket

  • If this is the first egress request for this researcher, create a new bucket in the Egress Project. Name it clearly so it is identifiable by researcher or workspace, for example: test-egress-WORKSPACENAME.

  • If the researcher already has a bucket from a previous request, you can reuse it — confirm it is empty before adding new files.

Grant access to the specific researcher

Important

Access must be granted to the specific requesting researcher only. Each researcher should only have visibility into their own egress bucket. If multiple researchers in the same environment have egress requests, each must have their own separate bucket.

Confirm files are in the egress bucket

  1. Log in to the Google Cloud Console and switch to the Egress Project.

  2. Go to Cloud Storage > Buckets and confirm the approved files are present in the egress bucket.

Initiate the transfer in Globus

  1. Go to app.globus.org and sign in with your Northwestern credentials.

  2. In the File Manager, search for your SDE egress collection by name. Northwestern IT will provide the exact collection name.

  3. In the left panel, navigate to the egress collection and select the files to transfer.

  4. In the right panel, select the destination collection provided by your recipient.

  5. Navigate to the destination folder and click Start.

Verify

  1. Globus will send an email notification when the transfer completes. Confirm with your recipient that the files arrived intact.

Attention

Delete the data from the egress bucket within 3 days of the transfer completing.

Common errors

Error

Plain-English meaning

What to do

Egress collection not found

The collection has not been provisioned or the name is wrong

Confirm with IT that setup is complete and get the exact collection name

Permission denied on destination

Your account is not authorized for the recipient’s collection

Ask your recipient to grant your Northwestern Globus identity access to their collection

Transfer failed or stalled

A network or endpoint issue interrupted the transfer

Globus will automatically retry; if it continues to fail, check the Globus troubleshooting guide