Closed pameyer closed 7 years ago
As a workaround, I ran count=0; while true; echo $((++count)); do /usr/local/dcm/gen/usg-wrap.bash; sleep 2; done
on the DCM so it would cook me an rsync script fast enough that it wasn't painful for me to work on integration (API) tests as I worked on https://github.com/IQSS/dataverse/issues/3352
Perhaps it's fine to use cron in a variety of places in the system but one place where I would really prefer for the user not to have to wait is when they are creating a dataset. I believe it leads to a much better user experience if they are able to download the rsync script right away after creating a dataset rather than waiting for cron to run.
There's still a glitch in the ansible/rq setup (https://github.com/sbgrid/data-capture-module/issues/13); but implementation for removing this worked w\ manual install.
Doesn't require API endpoint changes in the DCM.
dataverse has impedance mismatch when interacting with systems that depend on cron. needs re-factoring to eliminate this dependency, at least at the generation step. cron dependency for checksum validation less of a UI issue (there's already an async out-of-browser step); would also require more backend and transfer script changes.