This is Heather's DR2 timeline (to the best of Heather's knowledge, estimates for how long things will take). It started as a google document (here: ​https://docs.google.com/spreadsheets/d/11kmgFIT_d8k_bjsz4hhbUBjWsI_1trBl2LmWzruKcEo/edit#gid=446405522 , likely only editable by Heathers), but I wanted it on the wiki. I can't figure out how to make better looking tables in trac, so I made screen shots.

I have made a rough timeline for DR2 (more like an outline). The smallest unit of time in my timeline is 1 week, and I (arbitrarily) decided that Gene and I can be tasked on up to 3 tasks for DR2 at a time, Conrad gets 1 task at a time. Some tasks are clearly serial, (like, build, calibrate,test database), while others can be parallel (make P2 batches and test diff database). I created this mostly so I could have some guess as to how long things could take.

Conrad's numbers are pulled from either asking him, or looking up how long things took for DR1.

My numbers are estimated from either DR1 or other sources. There is overlap on different types of batches - this is because loading tapers off as a batch type nears completion, and it's possible to start up another batch type. I'm happy to fill available time (once dvo is calibrated) with loading whatever batches I can, but I do expect it to take quite a lot of weeks to make the batches.

Gene's numbers are guesses (I am happy to correct as necessary)

Not included is the testing/verification/investigations of ipptopsps/psps things, which is quite time consuming during loading, as well as all other Panstarrs related tasks.

V2: (5-5-2017) - we are still working on cluster issues. Of note, xnodes (which heather needs to make batches) are still unavailable due to power and networking constraints.

V3: (7-11-2017) - shifted Gene's items down; shuffled heather's items around to accommodate a different order of batch creation (FW still on track,but shifted due to cluster issues). Conrad's stuff shifted and reorganized (unexpectedly had to move his cluster, had cut point problems). As usual, not scheduled: unplanned downtime, vacations, etc.

V4: (11-30-2017) - shifted things down.

V5: (03-01-2018) - shifted things down, added in pole holes, etc.

Acronyms

IN - init batch - includes definitions of metadata tables, flags, etc.

P2 - Detection / ImageMeta tables

FW - ForcedWarp* tables

ST - Stack* tables

OB - ObjectThin/MeanObject tables

FO -ForcedObject* tables

FG/FFG ForcedGalaxy tables (I forgot exactly what they are called) - interchangably called FG/FFG - requires extensive work to put into DVO

DO - DiffObject tables - requires extensive work DF -DiffDetection tables - requires extensive work

• october - higher bandwidth for STSCI

• slice machine issues (need more details Ken/Conrad)