February 1, 2024

This release includes back-end updates to workflows, interactive analysis (Notebooks, Galaxy, RStudio), user interface, Data Repository, and Google and Azure integrations to improve upcoming features.

Data Repository

• Large field values (>8K bytes) are now supported in text and array fields in Azure datasets

Cloud Integrations

• We've added an additional spend category for Azure Billing Project Spend Reports to more accurately reflect workspace infrastructure costs. See this article on Terra expenses in the Azure portal for more information.
• Terra on Azure now supports more regions -- Central US, East US, South Africa North, South Central US, and West US 2 can now be selected when deploying the Terra Managed Application and all Terra resources will be created in the selected Azure region. South Central US is still the recommended region for using Terra.
• We resolved an issue where GCP workspaces could not be cloned from the workspaces list view.
• We resolved an issue where it was not possible to clone a Requester Pays GCP Workspace before a user billing project was selected. The billing project to bill during the clone process will be selected by the user within the Clone Workspace dialog. 

• Terra is adopting the modern `gcloud storage` command for localizing input files onto task VMs. All workflows using built-in localization will receive the enhancement automatically and do not need WDL updates. Delocalization (uploading) remains unchanged for now.

  With appropriately specified virtual machines, localization can now run at up to 1200 MB/s. Here are some guidelines for appropriately specified VMs:

• Tested with Intel Ice Lake CPU platform, 16 vCPU, 32 GB RAM, 2500 GB SSD. Runtime attributes used:

  runtime {
      docker: "rockylinux:9"
      disks: "local-disk 2500 SSD"
      cpu: 16
      memory: "32 GB"
      cpuPlatform: "Intel Ice Lake"
  }

  • Network throughput scales with vCPU count with a 1200 MB/s plateau at 16 vCPUs (For more information, see Google's documentation on VM configuration)

  • Disk throughput scales with disk size and type, reaching the 1200 MB/s VM limit with a 2.5 TB SSD. Worked example: 1200 MB/s ÷ 0.48 MB/s per GB = 2500 GB (For more information, see Google's documentation on VM storage configuration)