Project

General

Profile

13-June-2016

Attending: Bruno, Alex, Satish, Vladimir, Paul, Dan

Computing issues

dCache downtime on Thursday, probably all day. FTSs not working right now, a ticket already made by Alex. Probably a variable to include a proxy, but Alex wanted to confirm if Robert thinks this is the right approach. We confirmed Dan Persey as the new release manager, who will overlap with Paul for a few weeks.

Tags

prod2reco.g, some technical issues due to SuperNovaUtilities, fixed now. A new snapshot due by the end of the week. Adam generated some files with high intensity using latest prd2genie (for new fibre brightness). He couldn’t reconstruct due to missing headers (products missing). Newer versions should be backwards compatible, so pushing might not be the best solution. Instead, a new branch (renamed) with different tags. Paul will start working on this today. It’s going to require to re-generate Adam’s files. Samples got confused with one another. There will also be an ana branch upcoming probably on Friday. Paul’s making a list of tags not to be deleted (currently used). Soon, there’ll be a branch built using root 6, which will be a good opportunity for training the new person.

Samples

High intensity MC (Adam)

They're going to be retired due to issues

ND RHC (Enrique)

Alex helped Enrique to figure out the changes to our code necessary to set the right path for flux files in the OSG. These changes were done in stashcache.sh Basically the modifications introduced ask for the existence of the correct directory in the cluster. If the path exists, then PNSF_NOVA_DATA is set, if it does not exist, then the code remains unaltered. We carried the test at FNAL, and I think the test worked in the sense that we were able to finally set the correct path. I committed the changes, and I believe that we need a new cut in order to get the new version to work at CVMFS. Once this is available, I will proceed to test the changes in the OSG sites, for which I will use ~500 files from the ND RHC, as discussed in the past. The rest of the files, Vlad can proceed to reconstruct, but definitions aren’t available yet. Alex will cut a new version of NGU, even if sam for users is not ready.

FD MC with cosmic overlay (Vlad)

Finished on Friday, definitions ready.

Special cosmic processing, new fibre brightness (Cristiana)

Ready to start simulation.

System improvements

Metadata defaults (Cristiana / Joseph)

◦ Job-specific metadata (Adam)
◦ Declaring files after copying (Vlad)
◦ Stascache (Enrique)
◦ FTS error cataloguing (Satish): It needs 2 external libraries, including MatPlotLib, so Satish made a request to install one of them. The MatPlotLib can simply be installed in a novagpvm because it’s small.