Project

General

Profile

8/28/2014

Minutes taken by M.Tamsett & C.Group, docdb

Attendance

Matthew, Ryan, Chris, Gavin, Eric, Pavan, Kanika, Dominick, Craig, Jeny, Paola …

Personnel for first analysis (Matthew)

  • Nate: MC simulation
  • Gavin/Satish/Bruno: MC RECO/PID/CAF
  • Dominick: Data RECO/PID/CAF
  • Chris: LEM

Raw2root hand over progress (Adam, Jeny, Nate, Paola)

Paola and Jeny are ready to go, will meet with Adam for final validation then transition over next week.

Implementing keep up for reconstruction (All)

Conclusion from the conveners meeting last week and the 3pm meeting this week was that this is something we want in place ASAP. How exactly this works wasn’t defined then. At this meeting we came up with the plan that we’d set a light weight single step reco in place which runs in keep-up over the FD & ND data. This will run RECO & CAF but not PID. PID may be included later on if the need is demonstrated. The final CAF product will be the blinded files. Plan is not to switch releases on this on a timescale of less than a month.

Action Item: Dominick will put together the fcl for this and then work with Jeny and Paola to get this up and running in a keep up mode. He will also demonstrate to them the staged processing needed to back process all the old data without swamping the system.

Action Item: Add into a nova special field for this. Also think about including it in the file name. Matthew will follow this up with Adam.

Including MRCC/E in production (Kanika)

MR CC is a special stage of reconstruction that runs after PID to produce muon removed files. These files then must run through reconstruction, PID and CAF again. Will only need to be run on ND NuMI and ND genie files to start with. FHiCL files have been produced by Kanika and are ready to go.

Action Item: New data tiers and FTS configurations need to be defined for the new MRCC steps. Kanika is on shift with Adam next week so they can try and do this then.

Discussion of Nu_e need to reprocess PID (All)

The situation on what is needed here and when is needed isn’t total clear. This will be followed up on by Tricia and Mayly.

Misaligned geometry & simultaneous genie+cry reco progress (Gavin)

Gavin pointed out that the SIM tag used to produce these files is not consistent with the baseline files.

Action Item: Matthew will follow up with Gavin and Nate to reproduce a baseline sample to compare these against.

Names

File names are “unifified” on the grid. Is this the best plan?

Action Item: Matthew will start an email thread on this.

Update on Metadata issues (Adam, Gavin)

Adam wasn’t around today so we’ll follow this up in the future.

Disk space (Pavan, Craig, Gavin)

We are in better shape than we have been in a long time:

/nova/data      140T   94T   47T  67% /nova/data
/nova/ana        95T   76T   20T  80% /nova/ana
/nova/prod      100T   82T   19T  82% /nova/prod

Will keep pushing to free up this resource!

Action Item: Pavan will archive the NDOS data on /nova/data before it is purged.

Grid site ping job (Eric, Nate)

Eric ran this but forgot to turn off directory hashing so didn’t get any output. The results were as he expected. He noted that copy back is an issue with some (3?) sites (particularly SMU).

Action Item: Matthew and Craig will follow up with OSG to give them a heads up about our forthcoming production.

Monitoring updates (Matthew, Dominick, Eric)

The plan is to get summary information from SAMs jobs by combining Eric’s scripts and Dominick’s scripts.

AOB

  • No meeting on Monday due to a holiday in the USA.