You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When running local studio experiments via the completion service, or distributed experiments whose results get posted to S3 (or S3 emulator), over time studio just leaves a bunch of turds in whatever file system it is using.
To be clear, we use the completion service as a means for Distributed Compute and treat our studio experiments more like cattle than pets.
To date we have implemented our own coded cleanup by specifying our own 'cleanup_bucket' flag to the database and storage structures in the studio_config YAML to compensate for studio not cleaning up after itself. But really, since it is studio which is creating these files, to be symmetrical, studio should really be the one to clean up these files as well, yet no attempt is being made here at all.
The text was updated successfully, but these errors were encountered:
When running local studio experiments via the completion service, or distributed experiments whose results get posted to S3 (or S3 emulator), over time studio just leaves a bunch of turds in whatever file system it is using.
To be clear, we use the completion service as a means for Distributed Compute and treat our studio experiments more like cattle than pets.
To date we have implemented our own coded cleanup by specifying our own 'cleanup_bucket' flag to the database and storage structures in the studio_config YAML to compensate for studio not cleaning up after itself. But really, since it is studio which is creating these files, to be symmetrical, studio should really be the one to clean up these files as well, yet no attempt is being made here at all.
The text was updated successfully, but these errors were encountered: