WLCG Monitoring consolidation meeting 22-11-13
- Alberto, Lionel, Luca, Eddie, Andrea, Pablo, Marian, Costin, Stefan, Valentina, Pedro, Maarten, Nicolo', Alessandro
- remote: David, Alessandra, Laura
- Apologies: Julia
Going through the doc:
- Pablo introduce the new structure of the document, with introduction, description, actions that will be used as plan
- Each responsible have to provide a time estimation for its taks
- People encouraged to express interest in tasks
- Task and plan will be translated to section 9
- Alberto: when the plan will be ready, it will be movd to Jira and use it to track progresses. Agreed.
- Alberto: from the doc comments, more information needed on processing/aggregation directly from transport layer. Pablo and Luca will discuss it offline.
Going through experiment feedback twiki:
Alberto: suggest to move to bug tracking to Jira, the same one will be used for the plan.
Pablo will send the link to Jira around. Pablo ask Alessandra to move the bug to the main tracker. Alessandra agreed.
Alberto: in the list of important features, WLCG report/availability should be mentioned.
Maarten/Andrea: if there is common agreement on the need of SUM-like functionality, why do not keep old SUM?
Pablo:: SUM lacks some functionality, such as multiple FQAN support, which SSB provides. Proposal: next meeting will be dedicated to SUM investigation. Ivan will present plot replicating SUM functionality, history plot, multiple fqan. If the experiment agree it is the good direction effort in SUM to SSB direction will be pursed, otherwise old SUM may be considered.
Andrea: another limitation is the size of the table, with header very long and a lot of scroll needed to get the overeall picture. the old SUM approach of legend with metrics name is was much more effective.
Maarten: metrics name may contain useful information.
Andrea: About Condor-G probe plan?
Pablo: Associate is working on condor-g to get understanding on how the submission works.
Luca: we aim at re-using most of the code from the existing probes, but this much depend on the result of the investigation.
Maarten: any idea of a schedule?
Marian: it has been already announced at MB will be ready for next spring 2014
Pablo: Condor-g probes will need the name of the condor-g queues. This information should be from vofeed.
Alessandra: it should not be a problem for ATLAS.
Andrea: For CMS an option is to take information from bdii. ( and replicate WMS behaviour using ranks)
Pablo: Rather than having the probe depending on bdii, I would prefer to get info from bdii and put in vofeed.
Luca: CreamCE probe with payload support is in the validation phase. Currently only on lhcb dev box, it will be installed on lhcb and alice boxes.
Maarten: also Alice interested in the new CE probe
Next meeting will be about the SUM like interface and review on the report
Pablo: about document comments, good to have it by the end of next week.