Reviewed the Minerva meeting that happened the other day.
We need to have a "what's going on right now" dashboard. We don't want clutter from three weeks ago in our monitoring.
Output files of a typical job go in a FTS drop box.
Was a job really completed according to the experiments criteria?
Campaigns are for massive things like reprocessing.
Slicing a dataset into smaller pieces, and then having multiple jobs for the smaller pieces instead of one massive job.
HTCondor is like the home station for job submission; it submits the job and keeps track of exit codes, and whatever else. It is a big batch system.
Jobsub is just a wrapper around HTCondor.
Consumer status is just good or bad in sam.
A strategy for this project is to keep this system high level and delegate to other systems for details.
The team feels that we have a good direction for what this system should do. Meetings with other experiments will most likely solidify our direction. It's time that we start wrapping up the requirements document, and begin development.