MarcoWR0827
From: Mon 6/30/08
Monday
- UofC ATLAS meeting (3rd HEP)
- ATLAS shift
- Problems with production proxy probably due to CERN VOMS upgrade (then downgraded): order of extended attributes is different, need to use
-order
option
- performed some GUMS tests to asses the error: voms-gums-test.txt
- there seems to be a problem with the callout that always maps to
usatlas3
. Is it using different information? Is it for legacy/compatibility reasons?
- troubleshooting Fortran library (libgfortran) problem (error in tmp file creation):
At line 209 of file ../i686-slc3-gcc323-opt/loadamdcagddfromstring.f
Fortran runtime error: No such file or directory
-
-
- some docs that may be useful
FORTRAN77
http://www.star.le.ac.uk/~cgp/prof77.html#tth_sEc10.13
http://gcc.gnu.org/onlinedocs/gcc-3.4.6/g77/News.html
In EGCS 1.1 versus EGCS 1.0.3:
* Use tempnam, if available, to open scratch files (as in `OPEN(STATUS='SCRATCH')') so that the TMPDIR environment variable, if present, is used.
http://www.opengroup.org/onlinepubs/000095399/functions/tempnam.html
http://www.opengroup.org/onlinepubs/000095399/basedefs/stdio.h.html
http://www.star.le.ac.uk/~cgp/prof77.html#tth_sEc10.1
-
dq2-XXX
troubleshooting
- lcg-cp and LBNL srm-cp work
- Fermi srmcp fails (dq2-get hangs)
Tuesday
- ATLAS shift meeting (phone CERN)
- Add Jabber/Chat instructions to the twiki page
- TAG developers meeting (phone CERN)
- Working FDR2 Pathena by tomorrow evening
- MWT2 meeting (RI405) - skipped
- ATLAS shift
-
dq2-XXX
troubleshooting
- Monica's dataset registration is all messed up: PFNs are incorrect. How this happened? I was able to copy out those files in the past!
- court 12-2.30pm
Wednesday
- ATLAS DB meeting (phone CERN) - skipped
- ATLAS facilities meeting (phone EsNet from 405)
- defining some bugs of problems during the shift:
- Fixing monica's dataset registration (remove extra 'uct2-dc1.uchicago.edu')
- select to a file
- modify the file to prepare update
- mysqldump for backup
- execute update
f = open('marco-monica-in.txt')
fout = open('marco-monica-out.txt', 'w')
flin = f.readline()
outmask = 'update t_pfn set pfname="%s" where guid="%s" and pfname="%s";'
while flin:
#print 'LINE:', flin
ltok = flin.split()
if len(ltok)!=2:
print "Wrong format:", ltok
continue
#print ltok
oldnametok = ltok[0].split('uct2-dc1.uchicago.eduuser')
#print 'OLDNT:', oldnametok
newpfname = "%suser%s" % (oldnametok[0], oldnametok[1])
#print 'NEWPFN:', newpfname
outl = outmask % (newpfname, ltok[1], ltok[0])
#print outl
fout.write("%s\n" % outl)
#break
flin = f.readline()
fout.close()
f.close()
- Google talk instructions on shift page: https://twiki.cern.ch/twiki/bin/view/Atlas/ADCoS#IM_Instant_Messaging_for_shift_c
-
dq2-XXX
troubleshooting
- utilities work:
dq2-get -L UCT3 -s MWT2_UC -f user.MonicaDunford.mSugraGrid_25x25_m0_180_m12_330_A0_0_tanBeta_10_sgnMu_p_mTop_175_12.0.6.3_0_aod.pool.root user.MonicaDunford.mSugraGrid_25x25_v2
- looking for wrapper to isolate env variation
- change start line in dq2 executables
- Python wrepper (checking env and executable)
- uploaded key to access Panda machines at BNL
Thursday
-
srmcp
bug reported
- FDR2 Pathena example
- ran succesfully at BNL, failed at other 2 CEs
- nothing produced?
Friday
--
MarcoMambelli - 01 Jul 2008