1 / 6

Analysis Model Migration Plan

Analysis Model Migration Plan. Accessing the Mini Converting to the New (dual-read) Micro Producing the Skims Cleaning up the Tag Beta Development. Accessing the Mini. Having a complete Mini is the biggest underlying change WRT the old model Resources must be split between Mini and Micro

niyati
Download Presentation

Analysis Model Migration Plan

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Analysis Model Migration Plan Accessing the Mini Converting to the New (dual-read) Micro Producing the Skims Cleaning up the Tag Beta Development

  2. Accessing the Mini • Having a complete Mini is the biggest underlying change WRT the old model • Resources must be split between Mini and Micro • Mini data will appear with the 12 series (~now!) • Provide adequate resources for the (few) early people accessing the Mini • Allocate a reasonable staging space at Tier-A sites • Turn on HPSS staging for these databases • What should non-HPSS Tier-A sites do? • What should non-Objy (RAL) sites do? • Continue BetaMini development in the 12 series • Efficiency improvements and bug fixes • Persistent Composites • BetaMini redesign in 13 + 14 series • Make Beta work naturally with the Mini • Improve staging tools in 13+ 14 series David Nathan Brown 2 CMWG2 Workshop 18-19 October 2002

  3. New (Dual-Read) Micros • The main migration issue is to get people to accept using New Micro instead of ntuples • Use Eric’s MicroK implementation to make this transition • Some development is needed to make this production-quality • Save Full Micro information in AllEvents collection • Migrate with the 12 series data • Convert the reprocessed data as it arrives • Keep both formats at first to test and compare • Set a date for switching the reprocessing to the new format • End of the year? • Discourage the generation of AWG ntuples on this data • Provide help converting macros to the new format • Reduce disk space available for this purpose • Refine the implementation in the 13 + 14 series • Reduce the content of AllEvents (if necessary) • Base the persistence on a (Kanga) Reduced Mini? David Nathan Brown 3 CMWG2 Workshop 18-19 October 2002

  4. Skims • Skims need more support in the New Model • Core development to support AWG customization • Manpower to increase the skim frequency • Resources to hold deep copy skims • convert AWG ‘ntuple’ space ? • Convert 12 series reprocessing skims to New Micro format when that becomes available • Objy skims are still needed for the Mini • Produce New Model skims early (March?) 2003 • Support deep copy or pointer copy • Support content reduction and customization • Support Mini or New Micro output • AWGs should validate the content and size of their skims • Code development in the 12 series? David Nathan Brown 4 CMWG2 Workshop 18-19 October 2002

  5. Tag • New Model Tag is a subset of the existing tag • Use the existing tag with 12 series data • Develop a reduced Tag in the 13 + 14 series • Jim Smith is already collecting data on Tag usage • Encourage user migration by disabling functions in 12 series which will not be available in 13 + 14 • Annoying messages -> job abort David Nathan Brown 5 CMWG2 Workshop 18-19 October 2002

  6. Beta Development • Not really a migration issue • Affects performance, not functionality • New Model will not really ‘work’ without it • Core redesign effort should have high priority • Leverage existing BetaMini redesign effort • Develop in 13 + 14 series • Organize AWGs to improve physics code • Provide a profiling service (user provides code, gets profile) • Organize training for code optimization • Develop in 13 + 14 series David Nathan Brown 6 CMWG2 Workshop 18-19 October 2002

More Related