WLCG Information System Evolution Task Force

Thursday 24 Jan 2019, 15:00 → 16:25 Europe/Zurich

513/R-068 (CERN)

Meeting to discuss the evolution of the WLCG Information System

Attended: Adrian, Andrew, Balazs, Stephan, Julia

 

Andrew presented implementation of CRR for VAC and VCycle which can be generated automatically. For resourcecapacity and number of logical CPUs both VAC and Vcycle show the maximum possible numbers considering the available pool. It stays the same independently of number of actually running VMs. However, with VAC , if hypervisors go down for long time, the pool of CPUs goes down accordingly. In this respect the numbers which are reported are show dynamic capacity.

Another feature of this approach is that in case of several CEs (ARC with VAC for example), one file is generated per CE, not per site. However, there is no double counting, so data can be aggregated by the consumer (for example CRIC).

 

Walked through latest CRR version. Comments are summarized in the googledoc.

There were no major new comments/corrections to the document. Comments discussed during the meeting have been applied to the document.

It was concluded that the agreed format can be suggested for the implementation by  the sites which would like to use CRR in place of BDII.

wlcg-cric presentation was postponed to the next meeting.

 

 

 

15:00 → 15:15 CRR VAC-Vcycle implementation

Speaker: Andrew McNab (University of Manchester)

20190124-mcnab-vacjson.pdf

15:15 → 15:55 Discussion of the latest CRR JSON structure proposal

CRR structure