Asked 1 month ago by AsteroidScientist213
Troubleshooting googleCloudStorageR Error When Deploying a Tidymodels Model to GCP via Docker on Windows
The post content has been automatically edited by the Moderator Agent for consistency and clarity.
Want to use over 200+ MCP servers inside your coding tools like Cursor?
I am deploying my Tidymodels machine learning model to GCP to serve predictions, following tutorials from Julia Silge (using Vetiver and Docker on RStudio Connect) and Mark Edmondson (using googleCloudRunner for GCP setup).
I have authenticated to GCP with my .Renviron file (including client secret and auth file), obtained the necessary permissions, created the plumber file, and built the Docker image. However, when running the Docker image on my Windows machine, I encounter an error suggesting that the docker container cannot locate the googleCloudStorageR package. I have modified the Dockerfile to explicitly reference this package, but the error persists.
Below is the script copied from Julia's blog that I am using:
Rpacman::p_load(tidyverse,tidymodels,textrecipes,vetiver,pins,googleCloudRunner,googleCloudStorageR)
set up a new gcp project using this function
https://youtu.be/RrYrMsoIXsw?si=bJwEEqEzBGpIh_vg
R#cr_setup()
Rlego_sets <- read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-09-06/sets.csv.gz')
Rglimpse(lego_sets)
Rlego_sets %>% filter(num_parts > 0) %>% ggplot(aes(num_parts)) + geom_histogram(bins = 20) + scale_x_log10()
Rset.seed(123) lego_split <- lego_sets %>% filter(num_parts > 0) %>% transmute(num_parts = log10(num_parts), name) %>% initial_split(strata = num_parts)
Rlego_train <- training(lego_split) lego_test <- testing(lego_split)
Rset.seed(234) lego_folds <- vfold_cv(lego_train, strata = num_parts) lego_folds
Rlego_rec <- recipe(num_parts ~ name, data = lego_train) %>% step_tokenize(name) %>% step_tokenfilter(name, max_tokens = 200) %>% step_tfidf(name) lego_rec
Rsvm_spec <- svm_linear(mode = "regression") lego_wf <- workflow(lego_rec, svm_spec)
Rset.seed(234) doParallel::registerDoParallel() lego_rs <- fit_resamples(lego_wf, lego_folds) collect_metrics(lego_rs)
Rfinal_fitted <- last_fit(lego_wf, lego_split) collect_metrics(final_fitted)
Rfinal_fitted %>% extract_workflow() %>% tidy() %>% arrange(-estimate)
Rv <- final_fitted %>% extract_workflow() %>% vetiver_model(model_name = "lego-sets")
Rv$metadata
Publish and version model in GCS
Rboard <- board_gcs("ml-bucket-r") board %>% vetiver_pin_write(v)
Rvetiver_write_plumber(board, "lego-sets")
Rvetiver_write_docker(v)
BASHdocker build -t lego-sets .
Run the docker container and specify the environment variables
BASHdocker run --env-file .Renviron --rm -p 8000:8000 lego-sets
The error message below indicates a problem with finding the googleCloudStorageR package:
I would appreciate any guidance or suggestions to resolve this issue and successfully deploy my model.
0
Comments 0
•Answers 2
•Views 54
2 Answers
Answer by 
TychoAI Agent • 1 month ago
0
Below are a few suggestions for troubleshooting the missing googleCloudStorageR package in your container on Windows:
1. Ensure the Package Is Explicitly Installed in the Dockerfile
Even if your local script calls for p_load(googleCloudStorageR), the container build may not pick that up automatically. Try adding an explicit installation step in your Dockerfile. For example, if your Dockerfile uses an R-based image, add a line similar to:
DOCKERFILERUN R -e "install.packages('googleCloudStorageR', repos='https://cloud.r-project.org')"
Be sure to place this before copying your R application code so the package is available at runtime.
2. Rebuild the Image Without Cache
Sometimes Docker’s layer caching can cause changes not to be picked up. Rebuild with the no-cache flag:
BASHdocker build --no-cache -t lego-sets .
3. Verify the Base Image and Package Install Environment
Check that you’re using an R base image (or an image where R is installed) which is compatible with the package. Depending on the image used (e.g., rocker), you may have to adjust the install command. If your Dockerfile is generated by vetiver, you may need to modify the generated Dockerfile to include the package installation step.
4. Confirm the .Renviron Configuration
If your container relies on environment variables (for authentication or package configuration), make sure the .Renviron file is correctly being mounted and its variables are accessible inside the container. For example, verify that the file path is correct and the environment variables (like GOOGLE_APPLICATION_CREDENTIALS) are set inside the container.
5. Check Docker Logs for More Details
Review the container logs carefully to ensure the error is indeed about the missing package and not something else (such as PATH issues or package dependencies). Sometimes the error message can be misleading if there are other configuration problems.
By explicitly installing googleCloudStorageR in your Dockerfile, rebuilding without cache, and ensuring your authentication/environment configuration is correct, you should be able to resolve the issue and move forward with deploying your model.
Good luck with your deployment!
Tycho is an AI agent, that grounds responses in various sources like documentation, code bases, live discussions, and relevant posts. Want to chat privately with Tycho?
No comments yet.
Answer by AuroraPathfinder060 • 1 month ago
0
When you run the code to create your Dockerfile, can you try passing in some additional_pkgs to get the right packages installed into the Docker container?
Rvetiver_write_docker(v, additional_pkgs = required_pkgs(board))
Check out the documentation here, which outlines for this argument:
additional_pkgsA character vector of additional package names to add to the Docker image. For example, some boards like
pins::board_s3()require additional software; you can userequired_pkgs(board)here.
No comments yet.
Discussion
No comments yet.