TRE Change Log

Owned by Aaron Jackson
Last updated: yesterday at 8:44 am
3 min read

Service Workbench

Service Workbench (SWB) is the platform our TRE is built around. It manages access to projects and handles the creation of workspaces.

  • planned Improvements to the Dashboard page after logging in.

  • Mar 12, 2024 Current Backend dependencies update and removal of “$0” cost pane - there is no expected downtime during this deployment

  • Jan 16, 2024 Fix an issue with username resolution for those using University of Dundee Office 365 as their authentication provider. Upgrade the backend for our egress checking application.

  • Nov 21, 2023 Fix an issue where workspaces from different projects can be used in another user’s project.

  • Nov 15, 2023 Upgrade our backend services to use a new version of Node.js runtime.

  • Jul 27, 2023 Performance improvements to our internal synchronisation tool.

  • Jul 24, 2023 Apply workspace limit to currently selected project rather than all workspaces in all projects. Additionally, remove the “Create Research Workspace” button from the Workspaces tab, ensuring new workspaces are linked to a study.

  • Jul 13, 2023 Fix an issue with SWB where there are more than 100 studies.

  • Oct 5, 2022 Sets a default value during SAML logins to make it easier to provision user accounts in SWB.

Shared Services

Our shared services include things like the SQL Server, package repositories and license servers. We periodically have to apply changes to this environment to ensure it’s kept up to date, safe and secure.

  • Nov 19, 2024 Increase size of SQL Server and enable creation of resource pools (these will not be used until later)

  • Nov 12, 2024 Current Alterations to SQL backup

  • Aug 13, 2024 Installing SSL certificates.

  • May 14, 2024 Deploy security improvements.

  • Apr 23, 2024 Deploy a more permanent fix to the issues related to accessing package repositories (conda-forge and CRAN).

  • Mar 19, 2024 Resize some storage volumes used by our shared SQL Server. This will result in a period of downtime during our usual maintenance window.

TRE Workspace Configurations

A workspace configuration defines the infrastructure behind the creation of your workspace, such as firewall exceptions, storage policies and the virtual machine you will be interacting with. Overtime, we make changes to these to improve usability or security posture. Most projects will be provided with either a Windows or Ubuntu workspace configuration. This may be customised for the requirements of the project, to include specific software not included in our standard offering. Since the number of workspace configurations grows over time, there is a limit on how long we can provide support for issues relating to that specific configuration.

The tables below list the version number for each workspace configuration, some description and support status.

Windows

Version Name

Description

Version Name

Description

Windows 2-v2.0.1 Unsupported

Apr 18, 2022

Windows 2-v3.0 Unsupported

Sep 13, 2022 Introduces support for our integration SWB data output workflow which is described in Requesting outputs (my work) out of the TRE. Workspaces of this version can no longer be created.

Windows 2-v3.1.2 Unsupported

Jan 30, 2023 Fixes some minor issues with mounting S3 study data buckets.

Windows 2-v3.1.3 Support ending

Jul 25, 2023 Minor bug fix to ensure that the study data is mounted as read-only.

Windows 2-v3.1.4 Support ending

Aug 15, 2023 Adds an option to disable EC2 hibernation - this is required for EC2 instances running Windows with more than 16GB of RAM.

Windows 2-v3.2.2 supported

Jul 10, 2024 Fixes an issue due to AWS deprecating a CloudFormation parameter.

Ubuntu Linux

This is the preferred environment for those working on deep learning related research. The environment can be provisioned with a GPU, or it can be used simply because you prefer working in an Ubuntu environment.

Version Name

Description

Version Name

Description

Ubuntu Linux 2-v2.0 Unsupported

Apr 7, 2022

Ubuntu Linux 2-v3.0 Support ending

Sep 13, 2022 Introduced support for our new data output workflow. No new workspaces using this configuration can be created.

Ubuntu Linux 2-v3.1 Support ending

Jan 18, 2023 Allows hibernation to be disabled preventing access to larger EC2 instances.

Ubuntu Linux 2-v3.2.2 supported

Jul 10, 2024 Fixes an issue due to AWS deprecating a CloudFormation parameter.

Slurm

Slurm is a scheduler used in high performance computing environments. Within the TRE, it allows a user to submit their job to many computers to run a task in parallel.

Version Name

Description and changes

Version Name

Description and changes

Slurm-v1.0.4 Unsupported

The first generally available workspace configuration for Slurm. Storage is provided using Amazon FSx Lustre file system, mounted across all nodes.

Slurm-v1.1.2 Unsupported

Improvements to autoscaling and enabling access to S3 mounts from the login node, allowing access of existing study and user data, as well as the ability to submit egress requests from the login node workspace.

Slurm-v1.1.3 Unsupported

Minor security improvements to the way tags are handled within the CloudFormation stack.

Slurm-v1.1.4 Support ending

Apr 5, 2023 Adds support for FSx OpenZFS, in addition to FSx Lustre. This provides cost savings for cases where the performance of Lustre is not required.

Slurm-v1.1.5 SUPPORTED

Sep 14, 2023 Adds a fallback option in case node termination fails. Compute node will check if it has registered with a login node and shutdown if not.

Slurm-v1.1.6SUPPORTED

Sep 21, 2023 Fixes an issue where provisioned compute nodes do not have permission to access S3 mounts.

SAS Viya 3.5

Currently only available to University of Dundee Staff and Students.

Version Name

Description and changes

Version Name

Description and changes

SAS Viya-v0.0.5 SUPPORTED

Jul 4, 2023 SAS Viya accessible via AppStream. Supports access to project, study and user S3 buckets.

SAS Viya-v0.0.6 SUPPORTED

Jul 11, 2024 Fixes an issue due to AWS deprecating a CloudFormation parameter.