Launching your Server

Select an Instance Size

The Server Options page lets you select a computing environment tailored to your needs, ensuring you have the right resources for your projects.

When selecting a server for your JupyterHub environment, it’s important to choose a server instance that meets your computational needs without overcommitting resources. In order to keep operational costs low, we recommend that you start with the smallest server option available, and only move to a larger instance when you hit performance barriers. Below are the available server options along with their specifications and recommended use cases.

Shared Small Server

The Shared Small server is a multi-user environment where up to four different users may be using the same machine, with resource allocation depending on their respective workloads. This is the most efficient compute configuration and is the recommended option for most users, especially for those starting out or working on less intensive tasks.

Use Cases:

• Initial development and testing

• Educational purposes

• Light to moderate data analysis

Small Server (Dedicated)

The Small server is a dedicated machine just for you. This option is suitable for users who need more computational power and memory for their projects. It ensures that you have consistent access to the allocated resources without competition from other users.

Use Cases:

• More intensive data analysis and processing

• Running medium-sized machine learning models

• Projects requiring higher memory capacity

Medium and Large Servers

Medium and Large servers for more intensive projects are available by request, on a case-by-case basis. To request a larger server for your project, please email data-help@earthscope.org.

Select an Image

Several pre-configured compute environments are available to choose from. Use the dropdowns below to select your initial configuration. Please see Envrionment Management for more details on customizing your ephemeral server instance after launch, or bringing your own custom image.

GeoLab

The default GeoLab image contains a variety of python packages to enable a broad range of geophyiscal data analysis. This image inherits all software in pangeo/pytorch_notebook, and includes additional support for EarthScope SDK/CLI and other geophysics tools. For a complete list of packages included in this image, see the GeoLab github.

Note: If you believe we’ve missed a well-loved geophysical data analysis tool, please let us know by filling out the Geolab Feedback Form!

R Studio

The RStudio image is a stock image provided by 2i2c. EarthScope does not currently provide support or geophysical package extensions for this image. \

Short Course Images

EarthScope supports several educational short courses and workshops hosted in GeoLab throughout the year. If you are using GeoLab as part of one of these workshops, select the image corresponding to your course.

Short Course Images are available on a temporary basis only and contain software configurations specific to the course; they are not intended for public use and should not be used for ongoing research or external projects.

In some cases, short courses may use a custom image that is not available in the dropdown. See Other - Bring Your Own Image, below.

You can learn more about EarthScope’s educational short course offerings here.

Other - Bring Your Own Image

GeoLab is compatible with many other custom compute environments that are configured to run in JupyterHub.

Custom images must be built docker containers that are available in a public image repository (e.g., aws ECR, ACR, dockerhub, quay.io)

Select ‘Other’ from the dropdown menu and specify the public URL of the container.

See Environment Management for more details on building your own image.

Stopping Your Server and Logging Out

For improved cost and operational efficiency, it is important to release the resources that you have claimed during your session by shutting down your server instance and logging out of your EarthScope account once you’re finished. This allows the hub to re-allocate those resources to other users and saves money by not charging for unused compute.

The full shutdown sequence is as follows:

1. Navigate to File –> Hub Control Panel (this will open in a new browser tab)

2. In the new tab, push the big red button that looks like this

3. Once the button disappears, your server instance will be stopped.

Note

If you are using a “Shared Small” instance for your session, this will not shutdown the server for other users.

1. Next, click Log Out

2. Finally, close all browser tabs for GeoLab before logging in again.

If you follow these steps, you will avoid any errors that might appear if you try to start GeoLab from the same browser tab.