Project

General

Profile

09/02/15

Attendance

Matthew, Jonathan, Ryan, Chris, Alex, Joseph, Michael, Jeny, Bruno, Satish, Craig, Nate, Dominick.

News

  • The beam group are generating new gSimple flux files with a lot more information in. This enables more detailed systematic studies. We should pick these up on our next generation cycle, which might require unfreezing of some packages - but what isn’t clear to me. There was some discussion about how such changes would fit into our current first-analysis model, the conclusion was that the people interested in this information can do it out of non-first ana pieces. It was noted that re-doing rock generation might help data/MC disagreements. It was also noted that full simulation rock + detector would be needed to re-determine the poisson mean for the overlays. This will likely be done by Michael B. or in communication with him. Gavin will talk to people.
  • A service desk request to change some of our metadata to ints has been opened up. The transition required cannot be undertaken transparently. The conclusion in the room was that database unavailability is preferred to inconsistency.
  • Reco keep-up requested with S15-02-05
  • New tag has been made for calibration processing
  • FTS log file archival ticket in
  • A discussion was started about whether there was any need for rapid-turn around that doesn’t depend on low level things? No-one in the room knew a good reason to do so now.
  • Matthew learned how the FTS configuration works and setup things for Gavin's calibration shifts. He will look into the dropbox strategy later in the week and will communicate with Robert I. about this.
  • Some concerns have been raised about whether the official datasets page scales. An alternative is to switch to the one Matthew maintains: http://nusoft.fnal.gov/nova/production/first_analysis/index.html. Comments are welcome
  • Craig noted that the implementation of user quotas doesn't stop rogue grid users from filling the disk as grid jobs don't run under your username. He would like someone to implement a fast DF based tool. Nate started working on this and can help someone get started.

Hadronic simulation

Nate (simulation) - FTF_BIC sample is done, another one is already done, he’ll send that around the details of this after the meeting, the two "Q" are stalled because of bugs. Nate will continue to work on these.

Satish (reconstruction) - he completed that FTF_BIC and baseline samples, although it emerged during the meeting that the bad-channels backport hadn't been propagated fully at the time. These will need to be remade.

Gavin (PID) - to start when the above is resolved.

There was some discussion about the use of offsite nodes and debugging them. It was decided that Nate should focus on physics samples for this and that this debugging should fall elsewhere (maybe to Paola).

FD post-shutdown reco

Joseph has completed this. He noted that one run failed. It appears that this is due to formatting of comments in the bad channels database and will need a version roll to fix. It was decided to leave this for now and to pick the dropped runs up next time. It was decided that production should not process cosmics yet as studies on these are still on-going.

Bruno will start working on PID of these files later this week.

Event times

The issue with event times in the latest MC was tracked down to a problem with timezones. This has been resolved using Joao's code, however Jon P. pointed out that a solution consistent with bad channels would be best. Jon will communicate this to DQ. It was also noted that it would be good to have a consistency check run on these times before production begins.

Bad channels web service issue update

There is no bug. One outstanding ND run has issues which Jon is still working on.

Microslice counting

My take away form Friday's DQ meeting was that this is promising but still not 100% understood. So no reprocessing will begin yet. A possible long term plan could be to run a test “rapid-turn-around” with ND data to validate this. Then we could re-raw to root FD data before the next version of the calibration is ready (~1 month).

Other things to watch, threshold handling, sag, uncalibrated channel handling, there are some plans for lots of this, we’ll watch this space.

AOB

  • Craig posted some draft slides for the SPPM review, docdb-12808. Please take a look and comment before his practice talk on Thursday. The actual talk is March 4th.