REBUS and VOFeed

Europe/Zurich
513/R-068 (CERN)

513/R-068

CERN

19
Show room on map
Description
Three topics to discuss: *Project plan *Update on the CondorG probes) *REBUS functionality (pledges)

Local: Maarten, Luca, Nicolo, Pablo, Simone, Alberto, Julia, Stefan, Eddie, Costin, Ale

Remote: David
Apologies: Lionel

Minutes taken by SImone

 

 

Action List review: all tasks which needed to be done by January are indeed done. Looking at the ones foreseen by February:

  • Simplify the Nagio probes framework” has been re-scoped to a proposal.

  • All other items are well understood, some already done, others in progress

  • Maarten about “Twiki or Drupal for documentation”: there is the rumor that Drupal 7 and 8 are completely incompatible and we have a lot of stuff in Drupal 7. Sharepoint in addition is taking pace again. Embarking ourselves in Drupal seems not a good idea if things are not clarified.

 

Rebus Functionality:
Pablo: it was agreed that REBUS is a critical application not only for experiment operations but also for experiment management, sites and WLCG management. We should get the opinions of all. REBUS information has been looked at and there would be no problem putting the information in SSB.
Alberto: one of the requirements will be that the interface of the “new” REBUS will be similar to the old one.
Pablo: we would like to change both the frontend and the backend to make it compatible with the new consolidation project.
Maarten: the “new” functionality here is the INSERT from experiments and sites (they need to add/update information).
Discussion: do we need in the dashboard the capability of interactive modification? Might be better to have simply the capability to re-inject the data for some period through the message bus. This would avoid inconsistencies downstream. It needs to be discussed further.
Pablo showed a prototype of REBUS information in SSB.
Maarten: it would be good if the new REBUS would have some hard cuts to avoid that the information filled are completely unrealistic (for example very high number of HEPSPECS/logical core). Sometimes the numbers might also look correct (they are not illogical) but they are not. In this case we go back to the discussion about evaluating HEPSPECs from experiments frameworks. When the SSB will become the main repository of the REBUS information, we have to decide if the information are fetched from the BDII or inserted by the site admins.
Nicolo’: we have to keep in mind that there are applications which fetch currently the information from REBUS in JOSN format and inject it in experiment services; so a compatible JSON should be provided for the new framework.

 

AGIS and VO Feed:
there are information which in the new SAM framework we would need and today are not in the VO Feed. For example, the queue names, the space tokens, the experiments site contacts (we are not sure the last one is needed). Some information are needed because we will use Condor for job submission. All those information are in AGIS for ATLAS and in some respect for CMS (from the Common Analysis Framework). We agreed that:

  • The new format can be JSON or XML, the important is that there is validation at the level of the format.

  • Several additional information are needed and should be added to the VO Feed. The new fields will probably be mostly optional, and we should agree on the format.

  • Pablo will send around a first draft of the extension of the VO Feed for comments.

 

AOBs:

  • Stefan: The GSR has been re-discussed in the discussion about Common Projects and the current technology is REBUS. Maarten: the way we fetch information can be re-worked if GSR is delivered.

 

  • Meet again in 2 weeks at 14:00. There will be a status report of the transition of the current SAM to Greece/Croatia/France, and a presentation about the new SAM interface. Alessandro will take minutes. Costin (as the last person arriving to this meeting), will take the next.

 

There are minutes attached to this event. Show them.