r3 - 10 Jun 2010 - 10:01:57 - RichardMountYou are here: TWiki >  AtlasSoftware Web > Minutes30Apr2010

Minutes30Apr2010 RAC Minutes, April 30, 2010


Members (*=present, #=apologies)

*Richard Mount, Kevin Black begin_of_the_skype_highlighting     end_of_the_skype_highlighting (Physics Forum Chair), #Jim Cochran (Analysis Support Manager), *Alexei Klimentov (ATLAS ADC), Ian Hinchliffe (Physics Advisor), Rik Yoshida (Tier3 Coordinator), *Michael Ernst (U.S. Facilities Manager), Rob Gardner (Integration Coordinator), *Kaushik De (U.S. Operations Manager), *Armen Vartapetian (U.S. Operations Deputy)

Ex-Officio: *Torre Wenaus, #Stephane Willocq, *Mike Tuts, Howard Gordon

Correction/approval of minutes of previous meetings

The not-yet-approved minutes of previous meetings were approved.

Report from the (4/29/2010) CREM Meeting (Richard, Alexei)

ESD distribution was discussed following the recent observation that BNL was, for a time, the only location where users could analyze the 2010 Data ESDs. The US had engaged in manual replication of ESDs to T2s to remedy the situation [in the RAC discussion Kaushik said this had been very effective - the T2s were full with a long list of jobs waiting to execute].

David Charlton proposed that the prescription on ESD distribution from the computing model - one copy at BNL, one at other T1s (mainly RAL) and one distributed over the T2s - be tried out in practice. There was obviously room to do this if the 14-fold replication of the AODs (each 1/7 the size of an ESD) were reduced. It was agreed to try this and to approach RAL to request that user analysis be allowed to run on their ESDs.

The issue of space for MC output was raised by Borut just before the CREM meeting. There was almost no space for new MC production. What old MC should be retained depended partly on the needs for pileup simulation, which depended on the unknown future luminosity at 7 TeV. Borut agreed to present a detailed breakdown and specific proposals for the deletion of old MC (with notice to the physics community to allow complaints or moving data to GroupDisk).

In the RAC discussion, Kaushik pointed out that it was a Physics Coordination responsibility to find out why AODs were insufficient.

Alexei noted that there were currently 7 copies of data ESDs. ESDs were the most popular datasets, but the data AODs were being accessed (the most popular AOD, data10_7TeV*merge*AOD*, was accessed by 19 users form 15 sites according to statistics provided after the CREM meeting). The computing model makes no provision for MC ESDs at T2s, but David Charlton now recommends one copy at T2s. This copy will be located at US T2s. RAL has only 5% of MC ESDs; BNL has 100%. The total size is 800TB, of which only about 10% is actually in use.

Alexei stressed the hopelessness of managing these data manually and making all decisions in committee. The process had to be (at least) semi automated.

Disk Space Usage (Armen and Kaushik)

Armen presented a table.

Michael pointed out that there was nearly 1 PB of unaccounted space at BNL. Some of this was unallocated, some was "dark data" - on disk but not in the catalog - that should be cleaned out, and some was in GroupDisk, ProdDisk etc. that were not in Armen's table.

Armen and Kaushik agreed to work on a more detailed summary that would include all SpaceTokens and estimates of dark data and unallocated space.

Exercising the new Group Production/Additional Production mechanisms (Kevin - not yet confirmed)

Kevin could not attend the meeting. Richard will contact him.

US Tier 2 ESD Distribution

The RAC formally endorsed the Data and MC distribution plans that had been circulated to the RAC and discussed and amended by email over the last week. In accordance with the RAC policy of not introducing delays in urgent operational decisions, these plans had already been forwarded to ADC.

AOB

  • The WLCG jamboree on Data Access and Data Management (Amsterdam, June 16-18) was discussed briefly. The ATLAS participation would be decided by the ATLAS management. It was expected that several RAC members would participate.
  • The RAC time slot will be 12 noon Fridays. Most of those who had problems with this time had said it would be OK after the end of April.

Action Items

  1. 4/30/2010: Richard, contact Kevin about the progress of Exotics MC production.
  2. 4/9/2010: Kevin, Create first version of a Twiki guiding US physicists on requesting Additional Production. (In progress)
  3. 3/26/2010: All but especially core team, Find time during the ADC Workshop next week to identify a point-of-contact for Valid US Regional Production (on hold pending a better definition of what this task should be following successful completion of some Exotics Additional Production).

About This Site

Please note that this site is a content mirror of the BNL US ATLAS TWiki. To edit the content of this page, click the Edit this page button at the top of the page and log in with your US ATLAS computing account name and password.


Attachments


pdf USATLAS_storage_100430.pdf (9.5K) | RichardMount, 03 May 2010 - 19:18 | US-ATLAS Disk Space Usage April 2010
pdf Data-ESD-T2.pdf (43.1K) | RichardMount, 04 May 2010 - 14:09 | Data-ESDs at US Tier 2s
pdf MC-ESD-T2.pdf (49.8K) | RichardMount, 04 May 2010 - 14:10 | MC ESDs at US Tier 2s
 
Powered by TWiki
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback