Skip to main content

Using a workstation for R&D

After you are logged into a workstation, there's no limit what you can do! To help point you in the right direction, this guide answers common questions about workstations.

Which operating system do workstations run?​

Currently, each Deep Origin software blueprint provides Ubuntu Linux. To use another operating system, contact customer support.

What is the "bench-user" user?​

We've created a single user for everyone who logs into a workstation. This user is called bench-user, with a home directory at /home/bench-user/. All users have sudo access without a password.

What data is persistent between stop/start cycles?​

The entire home directory (/home/bench-user/) will be saved when you stop and start a workstation. This includes packages installed with conda, pip, R, and Julia. Additionally, apt packages will be replicated when you stop and start a workstation. All files that are saved to locations outside of the home directory and not to cloud storage will be lost when a workstation stops.

Please see the data persistence page for more information.

Which compute cluster is my workstation located in?​

When you create workstations and storage drives, you choose the computer cluster in which they should be provisioned. You can determinate the location of a workstation by inspecting the URLs for its applications or SSH instructions. This locations of a workstation is important to know if you are pulling data from external cloud-based storage.

What types of storage are attached to a workstation?​

Workstations have two types of storage provisioned by default.

  1. The persistent home directory, mounted at /home/bench-user/. This is an AWS Elastic Block Store (EBS) volume. Users can control the size of this volume when they create adn edit workstations.
  2. The ephemeral root directory, which contains all other workstation data. This is also an EBS volume. Users cannot control the size of this volume.

In addition, users can mount shared storage drives to workstations and pull managed data into workstations.

How can I configure the software packages in my workstation?​

Workstation software blueprints provide conda, pip, apt, and R's install.packages() for package management. We recommend using conda to install packages whenever possible.

More information about installing tools is available here. For assistance installing packages, contact customer support.

Can I select the R version used in RStudio Server?​

Please see the page on managing services with s6.

How can I use other web-based applications in a workstation?​

You can use VS Code (web) to access web-based applications that are not available from the "Connect" menu for a workstation.

For example, you can use CELLxGENE in the single-cell genomics blueprint this way. First, launch CELLxGENE from a terminal in your workstation:

Launch CELLxGENE with an example dataset
conda run --name cellxgene cellxgene launch https://github.com/chanzuckerberg/cellxgene/raw/main/example-dataset/pbmc3k.h5ad

Second, use VS Code (web) to forward the port for CELLxGENE to your computer:

  1. Launch VS Code (web) from the connect menu for the workstation.
  2. Inside VS Code (web), type F1 to bring up the command palette.
  3. In the command palette, type "ports" and select the first option.
  4. Open the "Local Address" for CELLxGENE on port 5005.
  5. VS Code (web) will open a new browser tab with CELLxGENE.

Alternatively, you can directly access web apps by navigating your browser to https://code-server-{workstation-id}.{compute-cluster-id}.bench.deeporigin.io/proxy/{port}/, replacing {workstation-id} with the ID of your workstation, {compute-cluster-id} with the ID of the cluster of the workstation (e.g., us-west-2.aws), and {port} with the port for your application. If VS Code (web) does not automatically forward the port for your application, you can use the "Add Port" button within VS Code (web) to manually configure VS Code (web) to forward the port.

How can I view HTML-based reports in a workstation?​

Some computational biology tools, such as fastqc and multiqc, produce reports in HTML format. It is convenient to view these reports in a workstation without having to download the files to your local machine. To view these files, we recommend installing a simple HTTP server in your workstation, launching this HTTP server, and using VS Code (web) to connect to this HTTP server to view your files.

First, use a terminal in your workstation to install and run an HTTP server into your workstation. By default, this HTTP server will run on port 8081.

Install and run http-server
npx -y http-server

Second, follow the instructions in the previous section to use VS Code (web) to access this HTTP server through port 8081. Once you connect to your HTTP server, you will see a static directory structure that you can use to navigate and open your HTML files.