Minutes21May2010 - RAC core team minutes, May 21, 2010
Core Team Participation
Armen Vartapetian, Torre Wenaus
Report from Thursday's CREM meeting
The CREM agenda is at http://indico.cern.ch/conferenceDisplay.py?confId=94843
DESD and TAG replication was introduced by Alexei in response to requests from Kaushik and Eric. It was agreed that Alexei would upgrade his code to allow DESD and AOD distribution to be different. DESD and AOD distribution has been 10 copies at T1 and 15 at T2. For the near future, DESD distribution would be reduced to ~10 copies overall (which just happens to conform to the CM). Eric's request to stop TAG replication to French T2s was deferred pending an evaluation of the additional merging of TAGs that is now underway.
Unused RAW data deletion was requested by Alexei. It was agreed that disk replicas not used for reprocessing and taken before April1 should be deleted.
RD Schaeffer noted that the sim AOD had grown by about 100kB (ttbar) due to a change in trigger menu and by about 15kB due to the lowering of the track pt threshold from 500 MeV to 100 MeV. It was commented that these issues are not CREM responsibilities.
Borut talked about the attempted/planned deletion of n-2 simulation data (created in March). Analysis groups needed enough notice to copy the datasets they needed to GroupDisk. At Hans's' suggestion, Borut agreed to write a script to list the unsubscribed files and delete these first. There was general agreement that n-2 data would often be needed and that groups would need tools to manage increasing volumes of data on GroupDisk.
Progress with Additional/group production example
Richard reported that Kevin had succeeded in getting the Drell-Yan production requests into the production system. This had involved discovery that a particular spreadsheet template had to be used to send the request to Borut. Documenting the detailed steps was expected to be very valuable. Kaushik said that the US T2s were 'empty', so the DY production should run quickly.
US Operational Issues
Kaushik discussed the revised US data distribution plans. Discussion centered on the plan to use the ATLAS central distribution mechanism to send data to Wisconsin DataDisk and Illinois DataDisk. The RAC core team agreed that it was preferable to use the central distribution system in such cases. The sites would make the data accessible to ATLAS via DQ2. The sites should only be encouraged to offer production of analysis capacity to ATLAS if it was likely that this could be made cost effective. The central effort required to support ATLAS production and analysis at non-Tier1/2 sites in the US had up to now been significant, and any expansion of this support needed a sound justification.
Further thoughts on "How to publicize current dataset distribution policies and realities as they affect the US?"
Richard had circulated the URL of a first attempt at a googledocs spreadsheet capturing dataset distribution targets with a US focus. It was agreed that this was useful and nobody had a better idea than googledocs. Richard had inserted interesting information on file sizes and numbers of files, but all agreed that this information could not easily be kept up to date and should be dropped from the next version. Some way of making old versions of the spreadsheet would be required. The spreadsheet should be linked from the RAC Wiki page as soon as possible.
- 5/7/2010: Richard, Create a web page summarizing the dataset distribution targets in the US (In progress)
- 4/9/2010: Kevin, Create first version of a Twiki guiding US physicists on requesting Additional Production. (In progress)
- 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.