Initial Questions

• Question about space requirement for databases - 1 TB of space. This will be for calibration and Tags.
• If you host a TAG database, it should be updated by the system when files arrived.
• Lets not wait for trigger information to be included.
• Note - an action on the Tag database will return a file set, not a dataset.
• This has to do with how POOL works - a persistent pointer infrastructure only knows about file GUIDS.
• Glasgow group is working on this, how to map a list of events and files into the DQ2 structure (datasets). Then, do you have them locally. Then there are implications for the execution infrastructure. Can the execution be local? There are a variety of steps here.
• How to do trigger streams become clustered, coming from Tier0?

Workflows

• How does one populate a Tag database.
• How will people query this, and build jobs from it.

Action Items

1. Build a mysql database, with some (as yet input) to be populate. Won't get too hung up on the precise schema.
2. Jack will assess what will be available when.
3. Record queries possible, and those we anticipate.
4. Move the tags through a subscription.
5. Run a query, select events, run a job locally, identify those that we can't.
6. Split events into jobs, respecting file boundaries.
7. Extract various event content.

Meetings

• Every other week in the database meeting timeslot, Marco will be invited.
• See readiness wiki from Jack