abba enables users to submit programs for background execution, allowing you to continue work as usual and not concern yourself with the job being terminated when you end your interactive session. Furthermore, by combining abba with services like Posit Connect, you can schedule jobs to run on a routine basis, such as nightly. While services like Connect allow you to run your work within a designated R environment, abba expands these capabilities by allowing you to send your work to a server or cluster environment, such as Kubernetes, in a secure manner.
abba has three major components:
- End user job submission functions
- Cluster/server interface functions
- Server API template
The server API template serves as the driver for the actual submission of jobs where they need to execute. This is intentionally separated to ensure job submission can be done securely, without giving an end user too much power or access.
The cluster/server interface functions simplify the management of job submission into a cluster. Currently only Kubernetes is supported. These methods manage that process and are utilized on within the API template.
The end user job submission functions are what the typical abba user will utilize. These functions provide a simple interface to target what the end user must concern themselves with, such as the program being submitted, the resources necessary, the container environment to run within, and necessary mounts.
Using function in abba, users can send a program to run and retrieve the results.
abba_submit_and_get_log(
"/home/mike.stackhouse/repos/abba/test_programs/test_program.R"
)abba has some convenience functions to simplify the setup of a job that you intend to schedule.
create_batch_job()This creates an Rmd file in your local directory following a simple template.
abba also makes it simple to create the necessary plumber API when setting up within your own environment.
create_batch_api()If in need of support, contact support@atorusresearch.com