WLCG Accounting Task Force Meeting

Europe/Zurich
513/R-068 (CERN)

513/R-068

CERN

19
Show room on map

Attended: Julia, Alessandra, Alessandro, Miguel, Steve, Ivan

Maria, Pepe, Antonio, John excused

 

Julia reported that the proposed changes were presented at the GDB and MB meetings. No objections for the changes. The only concern was expressed regarding of introduction of the new terms (metric names) replacing normalized CPU/wallclock.

After discussion, it was decided to go with the new naming convention adding tool tips and notes in all possible places to make clear what it means. If there is a too strong negative reaction or a lot of confusion we can think how to provide more clarification or revert back in the worst case.

Went through the document with the proposed changes. No comments/corrections/suggestions from people attending the meeting. Ivan confirmed that most of the changes would be implemented by the end of the next week and soem are already in place. Julia asked him to advertise the new version asap , so that people could start checking, since we would like to make the switch to the new portal/reports beginning of the year and to start sending new version of the reports together with the current official ones already in December.

Julia asked whether anyone knows what does it mean applying efficiency factor to the metrics related to space accounting which is mentioned in the notes of the pdf version of the reports. Nobody knew and apparently it is wrong and should be removed. Julia will add this requirement to the document.

The document is ready for Ivan as a specification for performing necessary changes.

Related to the migration to the new set of reports and new version of portal is the status of the new T0 accounting in APEL. Miguel explained that certain delay in injection of the July data in APEL was caused by the fact that accounting portal processing uses the time stamp field which is not mandatory  in the smry records according to documentation ( time stamp of the latest job processed during a given month). Instead of using a month and a year fields which are included in the CERN smry reports, this non-mandatory field is used by the portal. Julia asked how difficult would be to add this field at the point when data is recorded to the portal DB. Ivan thinks that it rather has to be done by APEL. Has to be followed up with APEL developers, John and Ivan. Meanwhile in order to make CERN new accounting data to be digested by the accounting report, Miguel introduced necessary changes adding this field in his smry reports. Julia is going to check with John and Miguel the progress in injecting CERN data from July to November, since this is one of the important factors which could prevent us of switching to the T0-1 reports generated by the portal (rather than by REBUS withe manual intervention).

Last point discussed was the progress in chasing of the problematic sites. Alessandra explained that she provided a twiki to facilitate the task of site admins in debugging eventual problems with the accounting reports. It is linked from the task force twiki. There are some issues with REBUS which sometime does not provide correct info for converting of time to work and this has to be followed up probably through the IS evolution task force.

No AOBs

 

 

 

 

There are minutes attached to this event. Show them.
    • 15:00 15:40
      Discussion on implementation of the changes in the portal and reports 40m
    • 15:40 16:00
      Follow up on sites which might have problems reporting to APEL 20m

      The progress is tracked here:
      https://twiki.cern.ch/twiki/bin/view/LCG/FollowUpOnProblematicSites