Closed ilveroluca closed 6 years ago
I have a patch in my fork that shows why. Basically, the startup command in the current chart redirects stdout
(without stderr
) from this patched run_galaxy_config.sh
script to a file (config_db.log
). I think it would make life easier to let that output flow to the regular k8s pod log.
This commit removes the config_db.log from our start-up procedure. Previously, this file captured only stdout from run_galaxy_config.sh and in case of failure was copied to the galaxy pvc so that it could be read after the pod restart. Now everything is visible from the pod logs, if you can get it before k8s destroys them.
Requires some small changes to the helm chart.
Tested on minikube. Came up ok and I was able to run the multivariate workflow test.