custom cloud environment

July 31, 2021 05:15
12 comments

H! I want to use custom RStudio cloud environment and not sure whether it's doable or not. If I can, is there a specific base image I should use? Thanks!

Comments

12 comments

Jason Cerrato
- August 02, 2021 13:49
Hi Sehyun Oh,

You should be able to make a custom RStudio cloud environment by extending our RStudio base image, which you can find here: us.gcr.io/anvil-gcr-public/anvil-rstudio-bioconductor:3.13.2

If we can help with anything else, please let us know!

Kind regards,

Jason

0
Sehyun Oh
- August 02, 2021 16:33
Hi Jason,

I used us.gcr.io/anvil-gcr-public/anvil-rstudio-bioconductor:3.13.2 as a base image to build my own with additional packages in it. But whenever I tried to use that new image, starting cloud environment takes forever - it says downtime required, thinks really hard, but never finish. Is there any template Dockerfile? Can I use any container repository?

- Sehyun

0
Jason Cerrato
- August 02, 2021 17:58
Hi Sehyun,

Hmm, are you just trying to add more packages to the base image? If so, I would recommend using a startup script to install the packages you need. You can read more about startup scripts here: https://support.terra.bio/hc/en-us/articles/360058193872

This will be a much easier solution than creating a separate Docker!

Kind regards,

Jason

0
Sehyun Oh
- August 02, 2021 18:05
Hi Jason,

I tried the startup script already, but tried both package installation and data download, which didn't work. I'll try once again only with the package installation and let you know how it goes.

- Sehyun

0
Jason Cerrato
- August 02, 2021 18:08
Please do! I'm really curious to know which part isn't working and see if we can get it resolved.

Jason

0
Sehyun Oh
- August 02, 2021 20:42
It seems like I hit the timeout error with the startup script? I could make the startup script work by installing only one package. But when I actually tried to install all 96 packages I need, it failed with the below error. Any suggestion from here?
```
OCI runtime exec failed: exec failed: container_linux.go:380: starting container process caused: exec format error: unknown
```
0
Sehyun Oh
- August 02, 2021 20:44
Also, when I used the custom container built on us.gcr.io/anvil-gcr-public/anvil-rstudio-bioconductor:3.13.2 (quay.io/shbrief/gss_docker:anvil_rstudio), I got this error.
```
The request content was malformed: DecodingFailure at .toolDockerImage: invalid container image quay.io/shbrief/gss_docker:latest
```
0
Jason Cerrato
- August 02, 2021 20:51
Hi Sehyun,

Thanks for testing that and getting back to me. Would you be able to share your Dockerfile for quay.io/shbrief/gss_docker:anvil_rstudio so I can take a closer look? I can follow up in a private email thread if you prefer.

Kind regards,

Jason

0
Sehyun Oh
- August 02, 2021 21:02
Sure! I just sent the response to support@broadinstitute.zendesk.com. Let me know if I should use a different email address. Thanks!

0
Katie Saund
- January 03, 2022 23:27
Hi Jason and Sehyun,

Any success on getting a custom RStudio cloud environment based on the RStudio base image working? I'm trying to do the same task and get the same error as Sehyun:

The request content was malformed: DecodingFailure at .toolDockerImage: invalid container image...

Thanks!

0
Sehyun Oh
- Edited January 03, 2022 23:56
Hi Katie,

For my case, the problem was a registry - I saved my container image in Quay, which is recommended for workflows (Dockstore/ WDL) but not supported for custom cloud environments. Terra accepts Docker images stored in GCR, GHCR, or DockerHub only.

https://support.terra.bio/hc/en-us/articles/360037143432-Docker-tutorial-Custom-cloud-environments-for-Jupyter-notebooks

Hope this helps. Thanks!

- Sehyun

0
Katie Saund
- January 04, 2022 21:03
Thank you, Sehyun!

I was using the Google Artifact Registry and it wasn't working. I just switched over to the Google Container Registry and that solved my problem.

Thanks again,

Katie

0

Please sign in to leave a comment.