Skip to content

Latest commit

 

History

History
50 lines (41 loc) · 2.76 KB

gcp_vm_setup.md

File metadata and controls

50 lines (41 loc) · 2.76 KB

GCP VM Setup

This documentation details how to setup a standard GCP Debian VM to build and run bulk_fhir_fetch from source. It also details how to configure the VM service account to have the ability to upload to FHIR store.

  1. Create a new linux virtual machine. To ensure bulk_fhir_fetch running on this VM has permission to upload to FHIR store:

    • While creating the virtual machine, you will need to ensure that the default service account on the VM has access to write to the FHIR store. This setting can be found under "Identity and API access" when creating a VM and you can either set access per-API for FHIR store or set "Allow full access to all Cloud APIs." The "Allow default access" setting will not work, as it sets read-only access.
    • After creating the virtual machine, ensure the VM service account has "Healthcare FHIR Resource Editor" permission on the relevant datastore.
      • Your default service account should be of the form: [email protected].
      • To verify that that service account has the "Healthcare FHIR Resource Editor" permission, navigate to Google Cloud > Healthcare > Browser and select the dataset of interest/ create one if none exists. Select the datastore of interest/ create one if none exists. On the PERMISSIONS section (on the right hand side) expand the "Healthcare FHIR Resource Editor" tab and if you see your default service account then your permissions are setup properly.
      • If you do not see your service account, you can add it by clicking on ADD PRINCIPAL, adding your service account into the New principals section and selecting "Healthcare FHIR Resource Editor" in the Select a role section and clicking SAVE.

  2. Install git:

    sudo apt-get install git

  3. If you would like to build bulk_fhir_fetch from source, Go must be installed:

    1. On the Go download website right-click on the "Linux" button from the featured downloads and copy the link address.
    2. Install wget: sudo apt-get install wget
    3. Download the Go installer to the machine: wget <copied download link>
    4. Follow the instructions on the Go website to remove the previous install and unpack the newly downloaded one. The command will look something like rm -rf /usr/local/go && tar -C /usr/local -xzf <installer name>

  4. Clone the repository

    git clone https://github.com/google/bulk_fhir_tools

  5. Build bulk_fhir_fetch:

    cd bulk_fhir_tools
go build cmd/bulk_fhir_fetch/bulk_fhir_fetch.go