Transferring files

Warning

As with all connections to the clusters, if you are not using a wired ethernet connection in a University campus building then you will need to turn on the VPN.

To transfer files to/from the clusters you can:

  • Use a program that supports one or both of the SCP and SFTP protocols to copy/move files to/from your own machine or from a remote machine to the cluster.

  • Use a Research Storage fileshare as common storage directly accessible from your own machine and from the clusters.

  • Use a program like curl or wget to download files directly to the clusters.

  • Use a flight session on Stanage or interactive session on Bessemer, to open a Firefox browser and interactively download directly to clusters.

Hint

Downloading directly to the cluster may be 10x to 100x faster than doing a transfer from your local desktop or laptop (particularly if connecting remotely via VPN) as this will avoid using your local device’s internet connection which is likely a bottleneck. If you are able, you should make direct downloads to the cluster.

Transfers with SCP/SFTP

Secure copy protocol (SCP) is a protocol for securely transferring computer files between a local host and a remote host or between two remote hosts. It is based on the Secure Shell (SSH) protocol and the acronym typically refers to both the protocol and the command itself.

Secure File Transfer Protocol (SFTP) is also a file transfer protocol. It is based on the FTP protocol with included SSH security components.

Hint

If you need to move large files (e.g. larger than a gigabyte) from one remote machine to the cluster you should SSH in to the computer hosting the files and use scp or rsync to transfer over to the other directly as this will usually be quicker and more reliable.

If you cannot SSH into the remote machine, consider an alternative direct transfer method listed below.

Using SCP in the terminal

If your local machine has a terminal and the scp (“secure copy”) command is available you can use it to make transfers of files or folders.

Where below substitute $CLUSTER_NAME with stanage or bessemer and $USER with your cluster username.

You should be prompted for your Duo MFA credentials after entering your password. Request a push notification or enter your passcode.

To upload, you transfer from your local machine to the remote cluster:

scp /home/user/file.txt $USER@$CLUSTER_NAME.shef.ac.uk:/home/$USER/

To download, you transfer from the remote cluster to your local machine:

scp $USER@$CLUSTER_NAME.shef.ac.uk:/home/$USER/file.txt /home/user/

To copy a whole directory, we add the -r flag, for “recursive”

scp -r $USER@$CLUSTER_NAME.shef.ac.uk:/home/$USER/my_results /home/user/

Using Filezilla

FileZilla is a cross-platform client available for Windows, MacOS and Linux for downloading and uploading files to and from a remote computer.

Download and install the FileZilla client from https://filezilla-project.org. After installing and opening the program, there is a window with a file browser of your local system on the left hand side of the screen and when you connected to a cluster, your cluster files will appear on the right hand side.

To connect to the cluster, we’ll just need make a new site and enter our credentials in the General tab:

Caution

By default Filezilla will save profiles in plaintext on your machine. You must ensure you use a master password to encrypt these credentials by changing the settings as shown in these instructions.

You can create a new site by selecting file from top menu bar then site manager which will open a dialog similar to:

Screenshot of Filezilla site manager dialog.


After hitting the new site button you can enter your credentials in the general tab:

  • Host: sftp://$CLUSTER_NAME.shef.ac.uk (replace $CLUSTER_NAME with stanage or bessemer)

  • User: Your cluster username

  • Password: Your cluster password (leave blank and fill this interactively if on a shared machine.)

  • Port: (leave blank to use the default port)

  • Protocol: sftp

  • Logon Type: Interactive

In the transfer settings tab limit the number of simultaneous connections to 1.

Save these details as a profile and then connect. You should be prompted for your Duo MFA credentials. Request a push notification or enter your passcode. You will now see your remote files appear on the right hand side of the screen. This process can be repeated to save a profile for each cluster.

You can drag-and-drop files between the left (local) and right (remote) sides of the screen to transfer files.

Using rsync

As you become more familiar with transferring files, you may find that the scp is limited. The rsync utility provides advanced features for file transfer and is typically faster compared to both scp and sftp. It is a utility for efficiently transferring and synchronizing files between storage locations including networked computers by comparing the modification times and sizes of files. The utility is particularly useful as it can also resume failed or partial file transfers by using the --append-verify flag.

Many users find rsync is especially useful for transferring large and/or many files as well as creating synced backup folders.

Caution

It is easy to make mistakes with rsync and accidentally transfer files to the wrong location, sync in the wrong direction or otherwise accidentally overwrite files. To help you avoid this, you can first use the --dry-run flag for rsync to show you the changes it will make for a given command.

Note

Be cautious when specifying paths with or without trailing slashes. Ensure that you understand how rsync interprets these slashes to prevent unintended outcomes.

rsync Behaviour with Trailing Slashes

With Trailing Slash on Source Directory:

rsync -av /source/directory/ /destination/directory

  • When you use a trailing slash on the source directory it tells rsync to copy the contents of the source directory into the destination directory.

Without Trailing Slash on Source Directory:

rsync -av /source/directory /destination/directory

  • When you don’t use a trailing slash on the source directory it tells rsync to copy the source directory itself and its contents into the destination directory.

Trailing Slash on Destination Directory:

rsync -av /source/directory/ /destination/directory/

  • When you use a trailing slash on the destination directory it tells rsync to copy the source directory itself and its contents into the destination directory.

Without Trailing Slash on Destination Directory:

rsync -av /source/directory/ /destination/directory

  • When you don’t use a trailing slash on the destination directory it tells rsync to copy the contents of the source directory into the destination directory.

The rsync syntax is very similar to scp. To transfer to another computer with commonly used options, where below substitute $CLUSTER_NAME with stanage or bessemer and $USER with your cluster username. You should be prompted for your Duo MFA credentials after entering your password. Request a push notification or enter your passcode:

rsync -avzP /home/user/file.iso $USER@$CLUSTER_NAME.shef.ac.uk:/home/$USER/

The a (archive) option preserves file timestamps and permissions among other things; the v (verbose) option gives verbose output to help monitor the transfer; the z (compression) option compresses the file during transit to reduce size and transfer time; and the P (partial/progress) option preserves partially transferred files in case of an interruption and also displays the progress of the transfer.

To recursively copy a directory, we can use the same options:

rsync -avzP /home/user/isos/ $USER@$CLUSTER_NAME.shef.ac.uk:/home/$USER/

This will copy the local directory and its contents under the specified directory on the remote system. If the trailing slash is omitted on the destination path, a new directory corresponding to the transferred directory (isos in the example) will not be created, and the contents of the source directory will be copied directly into the destination directory.

As before with scp, to download from the cluster rather than upload simply reverse the source and destination:

rsync -avzP $USER@$CLUSTER_NAME.shef.ac.uk:/home/$USER/isos /home/user/

How to download files directly to the cluster

Downloading files directly to the cluster is usually the quickest and most efficient way of getting files onto the clusters. Using your home connection will be a significant speed bottleneck compared to large amounts of download bandwidth available on the clusters. Directly downloading to the cluster avoids this bottleneck!

Using Firefox Browser

Firefox browser can be used on both Stanage and Bessemer. This will allow you to interactively navigate the web, login to websites and download files as you would do locally.

Graphical desktop access to an interactive session can be achieved using Flight Desktop and TigerVNC . Once you have loaded the GUI desktop, open a terminal at the bottom of the screen and enter the command firefox, which will launch a browser.

Using wget / curl

One of the most efficient ways to download files to the clusters is to use either the curl or wget commands to download directly.

The syntax for these commands is as below:

Downloading with wget

wget https://software.github.io/program/files/myprogram.tar.gz

Downloading with curl

curl -O https://software.github.io/program/files/myprogram.tar.gz

Using Git

The Git software and same named command can be used to download or synchronise a remote Git repository onto the clusters. This can be achieved by setting up Git and/or simply cloning the repository you desire.

For example, cloning the source of the make software:

[user@login1 make-git]$ git clone https://git.savannah.gnu.org/git/make.git
Cloning into 'make'...
remote: Counting objects: 16331, done.
remote: Compressing objects: 100% (3434/3434), done.
remote: Total 16331 (delta 12822), reused 16331 (delta 12822)
Receiving objects: 100% (16331/16331), 5.07 MiB | 2.79 MiB/s, done.
Resolving deltas: 100% (12822/12822), done.

Git is installed on the clusters and can be used on any node and all commands such as push, pull etc… are supported.

Using lftp

Hint

It is recommended that you use an alternative method than lftp if possible. Using lftp in the command line interface should be a last resort as it is a little difficult / confusing to use.

lftp is a command-line program client for FTP, FTPS, FXP, HTTP, HTTPS, FISH, SFTP, BitTorrent, and FTP over HTTP proxy.

If you need to login to an FTP server to make a direct download to a cluster, you can use the lftp client.

Connecting with lftp

Caution

Where possible please connect with the ftps protocol as this protects your username and password from hackers performing man in the middle or sniffing attacks!

Connecting to an FTP server can be achieved as follows:

lftp ftps://ftp.remotehost.com

When this connection is successful an lftp prompt will appear as follows:

lftp ftp.remotehost.com:~>

At this stage you can now login after being prompted for your password as follows:

lftp ftp.remotehost.com:~> login username
Password:

At this stage directory listing and changing directory can be achieved using the ls and cd commands. By default these commands run on the remote server. To run these commands on the local machine simply prefix each command with an ! i.e. !ls and !cd.

The get (download) and put (upload) commands can also be used.

Downloading with lftp

To download a file use the get command as follows:

lftp username@ftp.remotehost.com/> get myfile.txt -o mydownloadedfile.txt

Uploading with lftp

To upload a file use the put command as follows:

lftp username@ftp.remotehost.com/> put myfile.txt -o myuploadedfile.txt