Run R recipe on container when using Dataiku custom R built-in environment
Hi all,
I want to run an R recipe on a Kubernetes container but I get an error that a library is missing.
This recipe is using the DSS built-in R code environment. What made me question if, to use containerized execution, I need to create an R code environment with the container I want to use and set it as the code environment for this recipe.
Am I right?
Answers
-
Hi,
If you run the R recipe locally using the built-in environment, do you get the same error? If so, then you are likely trying to use or import a library that is not part of the base environment. If so, then you would indeed need to create an appropriate code environment and install the relevant packages. Then, as you are using containerized execution, you will also need to make sure that the code environment is usable for containerized execution by following the steps described in our documentation here:
https://doc.dataiku.com/dss/latest/containers/code-envs.html
Best,
Andrew
-
Hi @ATsao
,That is the odd part. The R recipe was always run using the built-in R environment. And even now if I run it with the built-in code environments it runs without problems.
If I run it using the Kubernetes container I get the error:
"Error in R process: simpleError : no library trees found in 'lib.loc' "
-
Hi,
To help you further investigate the matter, could you please open a support ticket directly and include a job diagnosis (corresponding to when it's failing when using K8s and working when running locally)? You can generate a job diagnosis by following the steps described here:
https://doc.dataiku.com/dss/latest/troubleshooting/problems/job-fails.html#getting-a-job-diagnosis
Thanks,
Andrew
-
