Run a rake task
Run a Rake task on EKS
To run a rake task in Kubernetes, create a job request which specifies the deployment in which to run the task, and the task you want to run.
For more on how to use job requests, refer to the job request user guide
Working with CSVs on Kubernetes
Some of our legacy rake tasks require uploading a CSV file. This is a throwback to our previous Puppet-based infrastructure and should be phased out now that we’re on Kubernetes, as containers are meant to be immutable and ephemeral.
Running rake tasks that require a CSV is not supported in job requests. The correct workaround is to get temporarily
elevated privileges and
use kubectl exec to upload the file and then run the task.
It is not possible to use the job request system to run the job once the file is uploaded because jobs are not run in the same pods.
kubectl cp foo.csv $somepod:/tmp && kubectl exec $somepod -- rake name_of_task
$ kubectl get pods
# returns list of pods, including
# whitehall-admin-c4c7c957c-9q966
$ kubectl cp ~/Downloads/tag.csv whitehall-admin-c4c7c957c-9q966:/tmp
# copies the file
$ kubectl exec whitehall-admin-c4c7c957c-9q966 -- rake data_hygiene:bulk_update_organisation[/tmp/tag.csv]
# runs the rake task