BYU

Office of Research Computing

Globus FAQ

See also the extensive FAQs directly from globus.org at https://docs.globus.org/faq/

BYU Office of Research Computing Globus FAQ

What is Globus?

See the main ORC Globus page or https://www.globus.org for a full introduction. The synopsis is that Globus is a convenient way to transfer large files, or large datasets.

Who can I share files with?

You can share files with anyone that has a Globus account. Globus accounts are free.

Who can get a Globus account?

Anyone can sign up for a free Globus account. Logging into Globus can use BYU CAS authentication, CAS authentication at other universities worldwide, or even a Google account. We recommend that our users register for an account using their BYU CAS authentication as outlined in our getting started instructions.

Can I share files with people outside the university who do not have BYU ORC accounts?

Absolutely! This is one of the great features of Globus. Create a guest collection and add permissions for the Globus users that you want to grant access to.

Can I share a folder where collaborators can upload and download files?

Yes. Guest collections can be either read-only or read-write.

Is Globus secure?

The answer is yes, absolutely, but possibly no. Globus uses file permissions and encryption to protect your data, and the Office of Research Computing makes continuing efforts to keep data secure on our systems. All of that being said, if YOU share your home directory as a guest collection with anonymous world-readable and writable permissions, your home directory is no longer secure.

Is Globus a cloud app? Should I be concerned about file security?

Globus sends files directly between two endpoints using encryption, and only those endpoints have access to those files. For instance, if using Globus to transfer files from a BYU endpoint to a University of Utah endpoint, the contents of those files would not in any way be accessible to the employees or servers at globus.org. Only the accounts on the endpoints that have been granted access to those files will have any access to that data. However, when browsing files or initiating transfers of files through Globus, the metadata for those files IS sent through the globus.org infrastructure. This means that the employees who run the globus.org infrastructure potentially have access to that metadata, including the names and paths to the files that were transferred or browsed.

Most BYU ORC users will not need to be concerned about globus.org having access to file metadata unless one or more of these conditions apply:

  • A user or group needs to avoid leaking sensitive information to outside entities
  • A user or group stores sensitive information contained in file or folder names (see below)
  • A user or group does not or cannot trust third parties with sensitive file and directory names (keeping in mind that the third party is contractually bound to protect this information). If this is a concern, rename the relevant files or do not use Globus. Contact us if you have any questions.

A couple of examples of sensitive information saved in filenames and its associated path:

  • /home/userbob123/labresults/embarrassing_medical_condition/positive/SomePersonsFullName.pdf
  • /home/userjohn123/Chemistry444/students_who_failed_the_exam/Famous_Alumni_Full_Name.txt

Where does Globus store files?

Globus doesn't store files, it is simply a way of transferring files between two endpoints that have the Globus software installed. It is possible to share and transfer files to and from ORC home directories, groups, scratch, and most any file system within the Office of Research Computing.

What is my quota?

Globus does not store data; it moves it, like rsync or scp. The file systems receiving data through Globus will be subject to whatever quotas apply to that file system. Please be aware of this, especially if allowing outside collaborators to upload files in your own guest collections. A collaborator at another university uploading files to your guest collection could easily put your group or home folder over quota without your knowledge!

Can I use Globus to move files off campus?

You can use Globus to move files to ANY other Globus collection, whether it is on or off BYU campus. You can even use Globus to move files to your laptop using Globus Connect Personal.

Can I use Globus to move files between various groups and accounts within ORC?

Yes. You can use Globus to move files between your home space, groups you have access to, scratch, or any guest collection that another user may have created that you have been granted access to.

Can I use Globus to move files to college or department file servers?

There are two options to make this happen. If its just for your own files you need access to, you could install Globus Connect Personal on a machine that has access to the college or department files servers, such as your own desktop machine in your office, or perhaps a laptop. To allow general access to the college or department file servers through Globus for ALL users, you will need to convince your college or department to install a Globus Connect Server using the BYU ORC site license. A department or college server may also have significant performance gains over the use of Globus Connect Personal on your own machine. There are other reasons why your department may want their own server. Have them contact us.

Can I use Globus to move files to BOX, or other cloud providers?

Our site license does not include this functionality on our servers. If you have a lot of money to contribute to the cause, please contact us. (It's expensive.) If Globus Connect Personal can see your cloud storage on your own machine as a drive or folder, it should be able to move files to and from that space.

Can I use Globus to copy files from a laptop or desktop computer?

Globus Connect Personal is free software for everyone that can be run on a laptop or desktop computer to share files through Globus. It is available to download at https://www.globus.org/globus-connect-personal

Is there a Globus CLI or an API?

Yes, there is both a command line interface and an application programming interface for Globus. Authentication for these interfaces still requires 2FA and is handled very similarly to logging in using the web interface.

Can moving files with Globus be part of a job on the supercomputer?

Yes, it is possible to automate transferring files using either the Globus CLI or API as part of a job on the supercomputer.

Will Globus send me an email when my file transfer is complete?

Yes, this is an option. You can look at the status of current and past jobs on Globus.org under "Activity".

How can I view the current status of my transfer jobs?

You can view your active jobs on globus.org once logged in under "Activity"

Will my files continue to copy if I close Globus?

Globus works by submitting jobs, so you can close or logout of the https://globus.org website, and your jobs will continue to run. If you are transferring files on your laptop or desktop computer, then the Globus Connect Personal software will need to be running for those transfers to function.

How fast are Globus transfers?

There are many factors that will affect the speed of Globus transfers, but in general, Globus will be the fastest method for moving files. Under the hood, Globus uses GridFTP, which is a powerful engine to very efficiently move either large numbers of files, or very large files. Our Globus servers are well-connected to our network and filesystems, and should give comparable file and network performance to our interactive or cluster nodes, but with the advantage of the increased transfer capabilities of GridFTP. Large copy jobs may span multiple servers and network links, as well as multiple file storage paths.

Is Globus better than rsync or scp?

Globus has several advantages over other tools when moving large datasets. Globus uses GridFTP behind the scenes, which was designed specifically for efficiently moving both very large files, and large numbers of files. It will even split the transferring of files across multiple servers for increased throughput. You can simply schedule your job and walk away, allowing Globus to handle the complexity of the task. You can receive an email when your job is complete to report any issues if there were any, or to just let you know it's done.

Who outside of BYU uses Globus?

Most HPC facilities at both universities and national laboratories have Globus Endpoints for sharing files.

What BYU ORC file spaces are available through Globus?

Our Globus servers have the BYU ORC /home /grphome and /nobackup filespaces found on the RHEL9 image. That includes home spaces, groups, scratch, and other spaces on our clusters. For a short while, Globus will include legacy paths found on the old RHEL7 image such as /fslhome, /lustre, etc.

Where can I access Globus from?

Globus is typically accessed through Globus.org using most any modern web browser. A CLI and an API are also available if automating transfers is needed.

How do file permissions work in Globus?

Globus can only access files on your behalf that your account has access to.

How do I sort file listings such that directories are at the top with files following that?

Under the View menu at the top right of the file listing in Globus, select Type as the Sort by type.