The download and installation process for R Studio Server was documented very well on their website. But this link had useful information as well as a directory of the R/Ubuntu versions for download. The command listed downloaded a version of R incompatible with the Ubuntu version running on the VM. One thing which took some figuring out was that the command listed under the Install R on your VM heading. When we used Google Cloud’s smallest VM (which only had ~3GB RAM) there was a memory error when running S-MultiXcan. To get a sense of the size of the servers, they were initiated with 50GB of disk space, 6.5 GB of RAM, and 1 processor.

Which made 36 different servers, each with RStudio available on port 8787. Gcloud compute -project 'introstatgen' disks create 'qgt-$ do bash gcloud_init.sh $i done Here is what we used to spin up an array: $ cat gcloud_init.sh Most commands in the Google Cloud Console can be replicated in the command line, and just before creating a VM or Snapshot in the Console, you can find a link which gives you the analogous command. To spin up multiple VMs, we used Google’s command line tools.

Anyway, you won’t have to do all of that if you want to use the most current snapshot. Installing, compiling, configuring, uploading, and permissions-ing was the long part. Most of the time (and therefore most of this document) was spent on step 2.

Spin up a bunch of new VMs from this snapshot.

Configure the VM as an RStudio server with everything installed, downloaded, etc.

To set everything up, we had a basic workflow:

Here’s a guide about how we set up that server. The servers needed pre-loaded data, access to a terminal, pre-compiled binaries for torus and fastenloc, and the correct python/R/Linux environments to run all of our analyses. For the one-day seminar, we had a hands-on lab where we decided we needed to set up R Studio Servers.

Rstudio server google cloud