Original Savannah ticket 46390 reported by egeland on Wed Jan 28 11:30:20 2009.
FileDownload successfully pre-validates killed validation processes. From Stuart:
Hi
I am seeing some of my transfers are passing the pre-transfer
validation check when the files don't exist. I see things like:
Argument "signal 1" isn't numeric in numeric eq (==) at
/srv/localstage/phedex/sw/slc4_amd64_gcc345/cms/PHEDEX/PHEDEX_3_1_1/perl_lib/PHEDEX/File/Download/Agent.pm
line 596.
2009-01-28 10:04:09: FileDownload[13000]: xstats: task=63557384
file=27197347 from=T1_US_FNAL_Buffer to=T2_UK_London_IC priority=5
report-code=0 xfer-code=-2 size=2008858845 t-expire=1233158691.3496
t-assign=1233076598.71238 t-export=1233076621.51298
t-inxfer=1233076962.6841 t-xfer=-1 t-done=1233110374.81901
lfn=/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root
from-pfn=srm://cmssrm.fnal.gov:8443/srm/managerv2?SFN=/11/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root
to-pfn=srm://gfe02.hep.ph.ic.ac.uk:8443/srm/managerv2?SFN=/pnfs/hep.ph.ic.ac.uk/data/cms/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root
detail=(file validated before transfer attempt) validate=(Executing:
/srv/localstage/phedex/SITECONF/London/PhEDEx/FileDownloadSRMVerify
pre srm://gfe02.hep.ph.ic.ac.uk:8443/srm/managerv2?SFN=/pnfs/hep.ph.ic.ac.uk/data/cms/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root
2008858845 cksum:3801113012 0 ~~ Job exited with status code signal 1
(1) ~~ ) job-log=(no job)
So it looks like the exit code is actually a string? and according to
(*) if a string is used with the == operator it will evaluate to 0,
and so i validate the file.
This isn't happening for every file so I'm not quite sure what's
causing it - running the validate script by hand it seems ok.
-bash-3.00$ /srv/localstage/phedex/SITECONF/London/PhEDEx/FileDownloadSRMVerify
pre srm://gfe02.hep.ph.ic.ac.uk:8443/srm/managerv2?SFN=/pnfs/hep.ph.ic.ac.uk/data/cms/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root
2008858845 cksum:3801113012 0
file srm://gfe02.hep.ph.ic.ac.uk:8443/srm/managerv2?SFN=/pnfs/hep.ph.ic.ac.uk/data/cms/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root
doesn't seem to exist
-bash-3.00$ echo $?
1
(and it doesnt exist - so that's ok)
So i assume it may be some occasional problem, that i haven't tracked down yet.
Original Savannah ticket 46390 reported by egeland on Wed Jan 28 11:30:20 2009.
FileDownload successfully pre-validates killed validation processes. From Stuart:
Hi
I am seeing some of my transfers are passing the pre-transfer validation check when the files don't exist. I see things like:
Argument "signal 1" isn't numeric in numeric eq (==) at /srv/localstage/phedex/sw/slc4_amd64_gcc345/cms/PHEDEX/PHEDEX_3_1_1/perl_lib/PHEDEX/File/Download/Agent.pm line 596. 2009-01-28 10:04:09: FileDownload[13000]: xstats: task=63557384 file=27197347 from=T1_US_FNAL_Buffer to=T2_UK_London_IC priority=5 report-code=0 xfer-code=-2 size=2008858845 t-expire=1233158691.3496 t-assign=1233076598.71238 t-export=1233076621.51298 t-inxfer=1233076962.6841 t-xfer=-1 t-done=1233110374.81901 lfn=/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root from-pfn=srm://cmssrm.fnal.gov:8443/srm/managerv2?SFN=/11/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root to-pfn=srm://gfe02.hep.ph.ic.ac.uk:8443/srm/managerv2?SFN=/pnfs/hep.ph.ic.ac.uk/data/cms/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root detail=(file validated before transfer attempt) validate=(Executing: /srv/localstage/phedex/SITECONF/London/PhEDEx/FileDownloadSRMVerify pre srm://gfe02.hep.ph.ic.ac.uk:8443/srm/managerv2?SFN=/pnfs/hep.ph.ic.ac.uk/data/cms/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root 2008858845 cksum:3801113012 0 ~~ Job exited with status code signal 1 (1) ~~ ) job-log=(no job)
So it looks like the exit code is actually a string? and according to (*) if a string is used with the == operator it will evaluate to 0, and so i validate the file.
This isn't happening for every file so I'm not quite sure what's causing it - running the validate script by hand it seems ok. -bash-3.00$ /srv/localstage/phedex/SITECONF/London/PhEDEx/FileDownloadSRMVerify pre srm://gfe02.hep.ph.ic.ac.uk:8443/srm/managerv2?SFN=/pnfs/hep.ph.ic.ac.uk/data/cms/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root 2008858845 cksum:3801113012 0 file srm://gfe02.hep.ph.ic.ac.uk:8443/srm/managerv2?SFN=/pnfs/hep.ph.ic.ac.uk/data/cms/store/mc/Summer08/QCD_EMenriched_Pt30to80/GEN-SIM-RECO/IDEAL_V9_v1/0018/00DB5EA9-5AEC-DD11-8B39-001E6878F716.root doesn't seem to exist -bash-3.00$ echo $? 1
(and it doesnt exist - so that's ok)
So i assume it may be some occasional problem, that i haven't tracked down yet.
The download log is here if needed: http://www.hep.ph.ic.ac.uk/~cmsprod/download (note its a 38M file - sorry I don't rotate logs very often)
The result of this is inconsistencies in TMDB.
This is presumably fixed in CVS, needs testing and inclusion in next release.