PHENIX DATA Location at CCJ ( 2009.12.25 update )

The principle of the PHENIX data location at CCJ. This principle is subject to change.

Disk
Raid disk: User should use /usr/local/bin/rcpx (or /opt/ccj/bin/rcpx) to retrieve the data on these disks to a local disk of each Linux node on which your analysis program is running.


phenix official data

disk	data	contact	since
/ccj/w/data18	run2_v03_burn1	A.Kiyomichi	2002/7	removed 2005/6
/ccj/w/data18	run2_v03_embed	A.Kiyomichi	2002/9	removed 2005/6
/ccj/w/data13,data25	run2pp_v01_burn1	T.Chujo/A.Kiyomichi	2003/11	removed 2005/04
/ccj/w/data23	spin-ndst	K.Tanida/Y.Goto	2003/11	removed 2004/12 => see run3_pp_ndst
/ccj/w/data21 => data33(2004/12)	run3_electron_ndst	M. Togawa	2004/03	removed(migrated to run3dAu-ndst)
/ccj/w/data33,38 => data41 => data36	run3_pp_ndst	M. Kaneta	2004/12 (moved 2008/09)
/ccj/w/data13,21,30	run4-62GeV-ndst	M. Kaneta	2004/12	removed 2005/08
/ccj/w/data15,26,33 =>26,50 =>data36	run3dAu-ndst	F. Kajihara	2005/01(moved 2008/9)
/ccj/w/data35	run4AuAu-ndst	T. Isobe	2005/03	removed 2007/1
/ccj/w/data27	run4pp-ndst	H. Torii	2005/06	removed 2007/1
/ccj/w/data02a,data02b,	run5pp-ndst	H. Torii	2005/07	removed on 06/3/21 ->data54-56
data41-50,	run5pp-ndst	H. Torii	2005/07	removed 2007/2
/ccj/w/data01a->data21,28	run5pp-ndst	K. Aoki	2005/12 ~2006/03	removed 2007/2
/ccj/w/data55,56	run5pp-ndst	Y. Fukao	2006/05	removed 2007/2
/ccj/w/data54	run5pp-ndst	T. Horaguchi	2006/06	removed 2007/2
/local01/ (ap119-148)	run6pp_62GeV_agg	S. Yokkaichi	2006/09	*1)
/local01/ (ap119-148)	run5pp-ndst(CNT*)	S. Yokkaichi	2007/02	*1)
/ccj/w/data40	run5pp-ndst	K. Tanida	2007/1
/ccj/w/data43,44,45,46 =>data35	run5pp-ndst	K. Tanida	2007/2 (moved 2008/3)
/ccj/w/data31,32	run6pp-ndst	K. Tanida	2007/4
/ccj/w/data32,35	run6pp-ndst-tr	S.Yokkaichi	2008/6-
node local disk	ndst:run9,8,6,5	T. Nakamura	2009/4-	see here




• *1)file list of /local01/ on IBM machines(ap119-148): /ccj/u/spool/localdatalist/ibm/ibmdatalist.txt




HPSS

• access method to CCJ-HPSS
  '/opt/ccj/bin/cftp' can be used to check the directry structure and filename. 'cftp' is a command for user- level archiving as same as 'rftp' in RCF. It can be used from all linux nodes and ccjnfs10-13.
  As same as RCF, user cannot use 'cftp' to archive the large size data files. If you need to back up the data files, please consult phenix-ccj-admin_AT_rarfaxp.riken.go.jp.
  
  
• directory structure on CCJ-HPSS
  
  • Real Data:(almost same as RCF-HPSS)

    /home/phnxsink/(RunTerm)/(category)/(some structure)/(filename)
             example:
        	   /home/phnxsink/run5/eventdata/...
               /home/phnxsink/run2001/eventdata/...
    		    

  • DST and nDST :(almost same as RCF-HPSS) about run5 nDST.

    		    /home/phnxreco/(RunTerm)/...
    		    example:
    		    /home/phnxreco/run5/run5pp_v02CCJ_pro72/...
    		    /home/phnxreco/run2000/v02b/dsts/...
    		    

HPSS - run5 nDST


• run5 v02 production was conducted by Zhengyun You.
• Usually, users can use /home/phnxreco/run5/run5pp_v02CCJ_pro72(and 73)/AG_TAR as follows.

  (Zhengyn's comment)
  All vol.tar contains nDST are created during production (before Aug.15,2006)
  After all have been processed once, from Aug.15, I checked some kinds of
  errors that might happen in production for all nDSTs,
  ,reprocessed them again and aggregated output in vol_2.tar.
  So files in vol_2.tar are update ones and should replace the same files in
  vol.tar
  
  Since I could do production and aggregation for PRDF and generate 22 kinds
  of output (22 PWGs).
  Some kinds of nDST remains the same as previous one in production (vol.tar),
  like the 5 root files in vol067(_2)-EWG_ERT.tar.
  It is because some other kinds of output (e.x DST_EVT_ALl or CNT_ERT...)
  from the same PRDF may have errors.
  
  So we need to download vol.tar and vol_2.tar, untar vol.tar first, vol_2.tar
  second,it will overwrite previous files with the same name automatically..
  By the way, in pro72, there is one vol005 (only this one) has vol005_3-
  xxx.tar files, which are update ones, generated in second repair.
  NO such vol_3.tar in pro73.
  

There are directories as follows under /home/phnxreco/run5/.

• to be used

  run5pp_v02CCJ_pro72

  (processed by pro.72,run168000-175000)

  run5pp_v02CCJ_pro73

  (processed by pro.73, run175000-180000 & run174000-174300 for comparison with pro.72)

• test version

  run5pp_v02CCJ_pro71

  (early test, no longer used)

  run5pp_v02CCJ_pro72_test_060416

  (test, no longer used)

  run5pp_v02CCJ_pro73_2

  (process some runs with pro.73again to test randomness, no longer used, forget it)

And there are directories as follows under /home/phnxreco/run5/run5pp_v02CCJ_XXXX

to be used

AG_TAR

(untar files in AG, rearrange them by flavor (PWG_ERT, CNT_ERT..) and tar every 100 files)

others

AG

(aggregation of all kinds of ndst, from every 10 PRDF, to one file )

DST

(nothing here)

NDST

(tar 22 kinds of ndst from one PRDF to one .tar file)

dump_run05_production_c_05.tar

(Bad events info dumped from Database in pro.73)

run5pp_v02CCJ_pro73.tar

(log files generated in production, ~/r01/w/run5pp_v02CCJ_pro73)

run5pp_v02CCJ_pro73.tar.idx

(index of run5pp_v02CCJ_pro73.tar)

useful link

• PHENIX data production at CCJ
• PHENIX database on CCJ