| 1 |
slacapra |
1.1 |
from JobType import JobType
|
| 2 |
|
|
from crab_logger import Logger
|
| 3 |
|
|
from crab_exceptions import *
|
| 4 |
|
|
from crab_util import *
|
| 5 |
|
|
import common
|
| 6 |
gutsche |
1.3 |
import PsetManipulator
|
| 7 |
slacapra |
1.1 |
|
| 8 |
slacapra |
1.41 |
import DataDiscovery
|
| 9 |
gutsche |
1.66 |
import DataDiscovery_DBS2
|
| 10 |
slacapra |
1.41 |
import DataLocation
|
| 11 |
slacapra |
1.1 |
import Scram
|
| 12 |
|
|
|
| 13 |
slacapra |
1.63 |
import os, string, re, shutil
|
| 14 |
slacapra |
1.1 |
|
| 15 |
|
|
class Cmssw(JobType):
|
| 16 |
gutsche |
1.38 |
def __init__(self, cfg_params, ncjobs):
|
| 17 |
slacapra |
1.1 |
JobType.__init__(self, 'CMSSW')
|
| 18 |
|
|
common.logger.debug(3,'CMSSW::__init__')
|
| 19 |
|
|
|
| 20 |
gutsche |
1.3 |
# Marco.
|
| 21 |
|
|
self._params = {}
|
| 22 |
|
|
self.cfg_params = cfg_params
|
| 23 |
gutsche |
1.38 |
|
| 24 |
gutsche |
1.44 |
# number of jobs requested to be created, limit obj splitting
|
| 25 |
gutsche |
1.38 |
self.ncjobs = ncjobs
|
| 26 |
|
|
|
| 27 |
slacapra |
1.1 |
log = common.logger
|
| 28 |
|
|
|
| 29 |
|
|
self.scram = Scram.Scram(cfg_params)
|
| 30 |
|
|
self.additional_inbox_files = []
|
| 31 |
|
|
self.scriptExe = ''
|
| 32 |
|
|
self.executable = ''
|
| 33 |
|
|
self.tgz_name = 'default.tgz'
|
| 34 |
corvo |
1.56 |
self.scriptName = 'CMSSW.sh'
|
| 35 |
spiga |
1.42 |
self.pset = '' #scrip use case Da
|
| 36 |
|
|
self.datasetPath = '' #scrip use case Da
|
| 37 |
gutsche |
1.3 |
|
| 38 |
gutsche |
1.50 |
# set FJR file name
|
| 39 |
|
|
self.fjrFileName = 'crab_fjr.xml'
|
| 40 |
|
|
|
| 41 |
slacapra |
1.1 |
self.version = self.scram.getSWVersion()
|
| 42 |
slacapra |
1.55 |
common.taskDB.setDict('codeVersion',self.version)
|
| 43 |
gutsche |
1.5 |
self.setParam_('application', self.version)
|
| 44 |
slacapra |
1.47 |
|
| 45 |
slacapra |
1.1 |
### collect Data cards
|
| 46 |
gutsche |
1.66 |
|
| 47 |
|
|
## get DBS mode
|
| 48 |
|
|
self.use_dbs_2 = 0
|
| 49 |
|
|
try:
|
| 50 |
|
|
self.use_dbs_2 = int(self.cfg_params['CMSSW.use_dbs_2'])
|
| 51 |
|
|
except KeyError:
|
| 52 |
|
|
self.use_dbs_2 = 0
|
| 53 |
|
|
|
| 54 |
slacapra |
1.1 |
try:
|
| 55 |
slacapra |
1.9 |
tmp = cfg_params['CMSSW.datasetpath']
|
| 56 |
|
|
log.debug(6, "CMSSW::CMSSW(): datasetPath = "+tmp)
|
| 57 |
|
|
if string.lower(tmp)=='none':
|
| 58 |
|
|
self.datasetPath = None
|
| 59 |
slacapra |
1.21 |
self.selectNoInput = 1
|
| 60 |
slacapra |
1.9 |
else:
|
| 61 |
|
|
self.datasetPath = tmp
|
| 62 |
slacapra |
1.21 |
self.selectNoInput = 0
|
| 63 |
slacapra |
1.1 |
except KeyError:
|
| 64 |
gutsche |
1.3 |
msg = "Error: datasetpath not defined "
|
| 65 |
slacapra |
1.1 |
raise CrabException(msg)
|
| 66 |
gutsche |
1.5 |
|
| 67 |
|
|
# ML monitoring
|
| 68 |
|
|
# split dataset path style: /PreProdR3Minbias/SIM/GEN-SIM
|
| 69 |
slacapra |
1.9 |
if not self.datasetPath:
|
| 70 |
|
|
self.setParam_('dataset', 'None')
|
| 71 |
|
|
self.setParam_('owner', 'None')
|
| 72 |
|
|
else:
|
| 73 |
|
|
datasetpath_split = self.datasetPath.split("/")
|
| 74 |
|
|
self.setParam_('dataset', datasetpath_split[1])
|
| 75 |
|
|
self.setParam_('owner', datasetpath_split[-1])
|
| 76 |
|
|
|
| 77 |
gutsche |
1.8 |
self.setTaskid_()
|
| 78 |
|
|
self.setParam_('taskId', self.cfg_params['taskId'])
|
| 79 |
gutsche |
1.5 |
|
| 80 |
slacapra |
1.1 |
self.dataTiers = []
|
| 81 |
|
|
|
| 82 |
|
|
## now the application
|
| 83 |
|
|
try:
|
| 84 |
|
|
self.executable = cfg_params['CMSSW.executable']
|
| 85 |
gutsche |
1.5 |
self.setParam_('exe', self.executable)
|
| 86 |
slacapra |
1.1 |
log.debug(6, "CMSSW::CMSSW(): executable = "+self.executable)
|
| 87 |
|
|
msg = "Default executable cmsRun overridden. Switch to " + self.executable
|
| 88 |
|
|
log.debug(3,msg)
|
| 89 |
|
|
except KeyError:
|
| 90 |
|
|
self.executable = 'cmsRun'
|
| 91 |
gutsche |
1.5 |
self.setParam_('exe', self.executable)
|
| 92 |
slacapra |
1.1 |
msg = "User executable not defined. Use cmsRun"
|
| 93 |
|
|
log.debug(3,msg)
|
| 94 |
|
|
pass
|
| 95 |
|
|
|
| 96 |
|
|
try:
|
| 97 |
|
|
self.pset = cfg_params['CMSSW.pset']
|
| 98 |
|
|
log.debug(6, "Cmssw::Cmssw(): PSet file = "+self.pset)
|
| 99 |
spiga |
1.42 |
if self.pset.lower() != 'none' :
|
| 100 |
|
|
if (not os.path.exists(self.pset)):
|
| 101 |
|
|
raise CrabException("User defined PSet file "+self.pset+" does not exist")
|
| 102 |
|
|
else:
|
| 103 |
|
|
self.pset = None
|
| 104 |
slacapra |
1.1 |
except KeyError:
|
| 105 |
|
|
raise CrabException("PSet file missing. Cannot run cmsRun ")
|
| 106 |
|
|
|
| 107 |
|
|
# output files
|
| 108 |
slacapra |
1.53 |
## stuff which must be returned always via sandbox
|
| 109 |
|
|
self.output_file_sandbox = []
|
| 110 |
|
|
|
| 111 |
|
|
# add fjr report by default via sandbox
|
| 112 |
|
|
self.output_file_sandbox.append(self.fjrFileName)
|
| 113 |
|
|
|
| 114 |
|
|
# other output files to be returned via sandbox or copied to SE
|
| 115 |
slacapra |
1.1 |
try:
|
| 116 |
|
|
self.output_file = []
|
| 117 |
|
|
tmp = cfg_params['CMSSW.output_file']
|
| 118 |
|
|
if tmp != '':
|
| 119 |
|
|
tmpOutFiles = string.split(cfg_params['CMSSW.output_file'],',')
|
| 120 |
|
|
log.debug(7, 'cmssw::cmssw(): output files '+str(tmpOutFiles))
|
| 121 |
|
|
for tmp in tmpOutFiles:
|
| 122 |
|
|
tmp=string.strip(tmp)
|
| 123 |
|
|
self.output_file.append(tmp)
|
| 124 |
|
|
pass
|
| 125 |
|
|
else:
|
| 126 |
gutsche |
1.50 |
log.message("No output file defined: only stdout/err and the CRAB Framework Job Report will be available")
|
| 127 |
slacapra |
1.1 |
pass
|
| 128 |
|
|
pass
|
| 129 |
|
|
except KeyError:
|
| 130 |
gutsche |
1.50 |
log.message("No output file defined: only stdout/err and the CRAB Framework Job Report will be available")
|
| 131 |
slacapra |
1.1 |
pass
|
| 132 |
|
|
|
| 133 |
|
|
# script_exe file as additional file in inputSandbox
|
| 134 |
|
|
try:
|
| 135 |
slacapra |
1.10 |
self.scriptExe = cfg_params['USER.script_exe']
|
| 136 |
|
|
if self.scriptExe != '':
|
| 137 |
|
|
if not os.path.isfile(self.scriptExe):
|
| 138 |
slacapra |
1.64 |
msg ="ERROR. file "+self.scriptExe+" not found"
|
| 139 |
slacapra |
1.10 |
raise CrabException(msg)
|
| 140 |
spiga |
1.42 |
self.additional_inbox_files.append(string.strip(self.scriptExe))
|
| 141 |
slacapra |
1.1 |
except KeyError:
|
| 142 |
spiga |
1.42 |
self.scriptExe = ''
|
| 143 |
|
|
#CarlosDaniele
|
| 144 |
|
|
if self.datasetPath == None and self.pset == None and self.scriptExe == '' :
|
| 145 |
|
|
msg ="WARNING. script_exe not defined"
|
| 146 |
|
|
raise CrabException(msg)
|
| 147 |
|
|
|
| 148 |
slacapra |
1.1 |
## additional input files
|
| 149 |
|
|
try:
|
| 150 |
slacapra |
1.29 |
tmpAddFiles = string.split(cfg_params['USER.additional_input_files'],',')
|
| 151 |
slacapra |
1.64 |
common.logger.debug(5,"Additional input files: "+str(tmpAddFiles))
|
| 152 |
|
|
for tmpFile in tmpAddFiles:
|
| 153 |
|
|
tmpFile = string.strip(tmpFile)
|
| 154 |
|
|
if not os.path.exists(tmpFile):
|
| 155 |
|
|
raise CrabException("Additional input file not found: "+tmpFile)
|
| 156 |
slacapra |
1.45 |
pass
|
| 157 |
slacapra |
1.64 |
storedFile = common.work_space.shareDir()+ tmpFile
|
| 158 |
|
|
shutil.copyfile(tmpFile, storedFile)
|
| 159 |
|
|
self.additional_inbox_files.append(string.strip(storedFile))
|
| 160 |
slacapra |
1.1 |
pass
|
| 161 |
slacapra |
1.64 |
common.logger.debug(5,"Inbox files so far : "+str(self.additional_inbox_files))
|
| 162 |
slacapra |
1.1 |
pass
|
| 163 |
|
|
except KeyError:
|
| 164 |
|
|
pass
|
| 165 |
|
|
|
| 166 |
slacapra |
1.9 |
# files per job
|
| 167 |
slacapra |
1.1 |
try:
|
| 168 |
gutsche |
1.35 |
if (cfg_params['CMSSW.files_per_jobs']):
|
| 169 |
|
|
raise CrabException("files_per_jobs no longer supported. Quitting.")
|
| 170 |
gutsche |
1.3 |
except KeyError:
|
| 171 |
gutsche |
1.35 |
pass
|
| 172 |
gutsche |
1.3 |
|
| 173 |
slacapra |
1.9 |
## Events per job
|
| 174 |
gutsche |
1.3 |
try:
|
| 175 |
slacapra |
1.10 |
self.eventsPerJob =int( cfg_params['CMSSW.events_per_job'])
|
| 176 |
slacapra |
1.9 |
self.selectEventsPerJob = 1
|
| 177 |
gutsche |
1.3 |
except KeyError:
|
| 178 |
slacapra |
1.9 |
self.eventsPerJob = -1
|
| 179 |
|
|
self.selectEventsPerJob = 0
|
| 180 |
|
|
|
| 181 |
slacapra |
1.22 |
## number of jobs
|
| 182 |
|
|
try:
|
| 183 |
|
|
self.theNumberOfJobs =int( cfg_params['CMSSW.number_of_jobs'])
|
| 184 |
|
|
self.selectNumberOfJobs = 1
|
| 185 |
|
|
except KeyError:
|
| 186 |
|
|
self.theNumberOfJobs = 0
|
| 187 |
|
|
self.selectNumberOfJobs = 0
|
| 188 |
slacapra |
1.10 |
|
| 189 |
gutsche |
1.35 |
try:
|
| 190 |
|
|
self.total_number_of_events = int(cfg_params['CMSSW.total_number_of_events'])
|
| 191 |
|
|
self.selectTotalNumberEvents = 1
|
| 192 |
|
|
except KeyError:
|
| 193 |
|
|
self.total_number_of_events = 0
|
| 194 |
|
|
self.selectTotalNumberEvents = 0
|
| 195 |
|
|
|
| 196 |
spiga |
1.42 |
if self.pset != None: #CarlosDaniele
|
| 197 |
|
|
if ( (self.selectTotalNumberEvents + self.selectEventsPerJob + self.selectNumberOfJobs) != 2 ):
|
| 198 |
|
|
msg = 'Must define exactly two of total_number_of_events, events_per_job, or number_of_jobs.'
|
| 199 |
|
|
raise CrabException(msg)
|
| 200 |
|
|
else:
|
| 201 |
|
|
if (self.selectNumberOfJobs == 0):
|
| 202 |
|
|
msg = 'Must specify number_of_jobs.'
|
| 203 |
|
|
raise CrabException(msg)
|
| 204 |
gutsche |
1.35 |
|
| 205 |
slacapra |
1.22 |
## source seed for pythia
|
| 206 |
|
|
try:
|
| 207 |
|
|
self.sourceSeed = int(cfg_params['CMSSW.pythia_seed'])
|
| 208 |
|
|
except KeyError:
|
| 209 |
slacapra |
1.23 |
self.sourceSeed = None
|
| 210 |
|
|
common.logger.debug(5,"No seed given")
|
| 211 |
slacapra |
1.22 |
|
| 212 |
slacapra |
1.28 |
try:
|
| 213 |
|
|
self.sourceSeedVtx = int(cfg_params['CMSSW.vtx_seed'])
|
| 214 |
|
|
except KeyError:
|
| 215 |
|
|
self.sourceSeedVtx = None
|
| 216 |
|
|
common.logger.debug(5,"No vertex seed given")
|
| 217 |
spiga |
1.57 |
try:
|
| 218 |
|
|
self.firstRun = int(cfg_params['CMSSW.first_run'])
|
| 219 |
|
|
except KeyError:
|
| 220 |
|
|
self.firstRun = None
|
| 221 |
|
|
common.logger.debug(5,"No first run given")
|
| 222 |
spiga |
1.42 |
if self.pset != None: #CarlosDaniele
|
| 223 |
|
|
self.PsetEdit = PsetManipulator.PsetManipulator(self.pset) #Daniele Pset
|
| 224 |
gutsche |
1.3 |
|
| 225 |
slacapra |
1.1 |
#DBSDLS-start
|
| 226 |
|
|
## Initialize the variables that are extracted from DBS/DLS and needed in other places of the code
|
| 227 |
|
|
self.maxEvents=0 # max events available ( --> check the requested nb. of evts in Creator.py)
|
| 228 |
|
|
self.DBSPaths={} # all dbs paths requested ( --> input to the site local discovery script)
|
| 229 |
gutsche |
1.35 |
self.jobDestination=[] # Site destination(s) for each job (list of lists)
|
| 230 |
slacapra |
1.1 |
## Perform the data location and discovery (based on DBS/DLS)
|
| 231 |
slacapra |
1.9 |
## SL: Don't if NONE is specified as input (pythia use case)
|
| 232 |
gutsche |
1.35 |
blockSites = {}
|
| 233 |
slacapra |
1.9 |
if self.datasetPath:
|
| 234 |
gutsche |
1.35 |
blockSites = self.DataDiscoveryAndLocation(cfg_params)
|
| 235 |
slacapra |
1.1 |
#DBSDLS-end
|
| 236 |
|
|
|
| 237 |
|
|
self.tgzNameWithPath = self.getTarBall(self.executable)
|
| 238 |
slacapra |
1.10 |
|
| 239 |
slacapra |
1.9 |
## Select Splitting
|
| 240 |
spiga |
1.42 |
if self.selectNoInput:
|
| 241 |
|
|
if self.pset == None: #CarlosDaniele
|
| 242 |
|
|
self.jobSplittingForScript()
|
| 243 |
|
|
else:
|
| 244 |
|
|
self.jobSplittingNoInput()
|
| 245 |
corvo |
1.56 |
else:
|
| 246 |
|
|
self.jobSplittingByBlocks(blockSites)
|
| 247 |
gutsche |
1.5 |
|
| 248 |
slacapra |
1.22 |
# modify Pset
|
| 249 |
spiga |
1.42 |
if self.pset != None: #CarlosDaniele
|
| 250 |
|
|
try:
|
| 251 |
|
|
if (self.datasetPath): # standard job
|
| 252 |
|
|
# allow to processa a fraction of events in a file
|
| 253 |
|
|
self.PsetEdit.inputModule("INPUT")
|
| 254 |
|
|
self.PsetEdit.maxEvent("INPUTMAXEVENTS")
|
| 255 |
|
|
self.PsetEdit.skipEvent("INPUTSKIPEVENTS")
|
| 256 |
|
|
else: # pythia like job
|
| 257 |
|
|
self.PsetEdit.maxEvent(self.eventsPerJob)
|
| 258 |
spiga |
1.57 |
if (self.firstRun):
|
| 259 |
|
|
self.PsetEdit.pythiaFirstRun("INPUTFIRSTRUN") #First Run
|
| 260 |
spiga |
1.42 |
if (self.sourceSeed) :
|
| 261 |
|
|
self.PsetEdit.pythiaSeed("INPUT")
|
| 262 |
|
|
if (self.sourceSeedVtx) :
|
| 263 |
|
|
self.PsetEdit.pythiaSeedVtx("INPUTVTX")
|
| 264 |
gutsche |
1.50 |
# add FrameworkJobReport to parameter-set
|
| 265 |
|
|
self.PsetEdit.addCrabFJR(self.fjrFileName)
|
| 266 |
spiga |
1.42 |
self.PsetEdit.psetWriter(self.configFilename())
|
| 267 |
|
|
except:
|
| 268 |
|
|
msg='Error while manipuliating ParameterSet: exiting...'
|
| 269 |
|
|
raise CrabException(msg)
|
| 270 |
gutsche |
1.3 |
|
| 271 |
slacapra |
1.1 |
def DataDiscoveryAndLocation(self, cfg_params):
|
| 272 |
|
|
|
| 273 |
gutsche |
1.3 |
common.logger.debug(10,"CMSSW::DataDiscoveryAndLocation()")
|
| 274 |
|
|
|
| 275 |
|
|
datasetPath=self.datasetPath
|
| 276 |
|
|
|
| 277 |
slacapra |
1.1 |
## Contact the DBS
|
| 278 |
slacapra |
1.41 |
common.logger.message("Contacting DBS...")
|
| 279 |
slacapra |
1.1 |
try:
|
| 280 |
gutsche |
1.66 |
|
| 281 |
|
|
if self.use_dbs_2 == 1 :
|
| 282 |
|
|
self.pubdata=DataDiscovery_DBS2.DataDiscovery_DBS2(datasetPath, cfg_params)
|
| 283 |
|
|
else :
|
| 284 |
|
|
self.pubdata=DataDiscovery.DataDiscovery(datasetPath, cfg_params)
|
| 285 |
slacapra |
1.1 |
self.pubdata.fetchDBSInfo()
|
| 286 |
|
|
|
| 287 |
slacapra |
1.41 |
except DataDiscovery.NotExistingDatasetError, ex :
|
| 288 |
slacapra |
1.1 |
msg = 'ERROR ***: failed Data Discovery in DBS : %s'%ex.getErrorMessage()
|
| 289 |
|
|
raise CrabException(msg)
|
| 290 |
|
|
|
| 291 |
slacapra |
1.41 |
except DataDiscovery.NoDataTierinProvenanceError, ex :
|
| 292 |
slacapra |
1.1 |
msg = 'ERROR ***: failed Data Discovery in DBS : %s'%ex.getErrorMessage()
|
| 293 |
|
|
raise CrabException(msg)
|
| 294 |
slacapra |
1.41 |
except DataDiscovery.DataDiscoveryError, ex:
|
| 295 |
gutsche |
1.66 |
msg = 'ERROR ***: failed Data Discovery in DBS : %s'%ex.getErrorMessage()
|
| 296 |
slacapra |
1.1 |
raise CrabException(msg)
|
| 297 |
|
|
|
| 298 |
|
|
## get list of all required data in the form of dbs paths (dbs path = /dataset/datatier/owner)
|
| 299 |
gutsche |
1.3 |
common.logger.message("Required data are :"+self.datasetPath)
|
| 300 |
|
|
|
| 301 |
gutsche |
1.35 |
self.filesbyblock=self.pubdata.getFiles()
|
| 302 |
mkirn |
1.37 |
self.eventsbyblock=self.pubdata.getEventsPerBlock()
|
| 303 |
|
|
self.eventsbyfile=self.pubdata.getEventsPerFile()
|
| 304 |
gutsche |
1.3 |
|
| 305 |
slacapra |
1.1 |
## get max number of events
|
| 306 |
|
|
self.maxEvents=self.pubdata.getMaxEvents() ## self.maxEvents used in Creator.py
|
| 307 |
gutsche |
1.44 |
common.logger.message("The number of available events is %s\n"%self.maxEvents)
|
| 308 |
slacapra |
1.1 |
|
| 309 |
slacapra |
1.41 |
common.logger.message("Contacting DLS...")
|
| 310 |
slacapra |
1.1 |
## Contact the DLS and build a list of sites hosting the fileblocks
|
| 311 |
|
|
try:
|
| 312 |
slacapra |
1.41 |
dataloc=DataLocation.DataLocation(self.filesbyblock.keys(),cfg_params)
|
| 313 |
gutsche |
1.6 |
dataloc.fetchDLSInfo()
|
| 314 |
slacapra |
1.41 |
except DataLocation.DataLocationError , ex:
|
| 315 |
slacapra |
1.1 |
msg = 'ERROR ***: failed Data Location in DLS \n %s '%ex.getErrorMessage()
|
| 316 |
|
|
raise CrabException(msg)
|
| 317 |
|
|
|
| 318 |
|
|
|
| 319 |
gutsche |
1.35 |
sites = dataloc.getSites()
|
| 320 |
|
|
allSites = []
|
| 321 |
|
|
listSites = sites.values()
|
| 322 |
slacapra |
1.63 |
for listSite in listSites:
|
| 323 |
|
|
for oneSite in listSite:
|
| 324 |
gutsche |
1.35 |
allSites.append(oneSite)
|
| 325 |
|
|
allSites = self.uniquelist(allSites)
|
| 326 |
gutsche |
1.3 |
|
| 327 |
gutsche |
1.35 |
common.logger.message("Sites ("+str(len(allSites))+") hosting part/all of dataset: "+str(allSites))
|
| 328 |
|
|
common.logger.debug(6, "List of Sites: "+str(allSites))
|
| 329 |
|
|
return sites
|
| 330 |
gutsche |
1.3 |
|
| 331 |
gutsche |
1.35 |
def jobSplittingByBlocks(self, blockSites):
|
| 332 |
slacapra |
1.9 |
"""
|
| 333 |
gutsche |
1.35 |
Perform job splitting. Jobs run over an integer number of files
|
| 334 |
|
|
and no more than one block.
|
| 335 |
|
|
ARGUMENT: blockSites: dictionary with blocks as keys and list of host sites as values
|
| 336 |
|
|
REQUIRES: self.selectTotalNumberEvents, self.selectEventsPerJob, self.selectNumberofJobs,
|
| 337 |
|
|
self.total_number_of_events, self.eventsPerJob, self.theNumberOfJobs,
|
| 338 |
|
|
self.maxEvents, self.filesbyblock
|
| 339 |
|
|
SETS: self.jobDestination - Site destination(s) for each job (a list of lists)
|
| 340 |
|
|
self.total_number_of_jobs - Total # of jobs
|
| 341 |
|
|
self.list_of_args - File(s) job will run on (a list of lists)
|
| 342 |
|
|
"""
|
| 343 |
|
|
|
| 344 |
|
|
# ---- Handle the possible job splitting configurations ---- #
|
| 345 |
|
|
if (self.selectTotalNumberEvents):
|
| 346 |
|
|
totalEventsRequested = self.total_number_of_events
|
| 347 |
|
|
if (self.selectEventsPerJob):
|
| 348 |
|
|
eventsPerJobRequested = self.eventsPerJob
|
| 349 |
|
|
if (self.selectNumberOfJobs):
|
| 350 |
|
|
totalEventsRequested = self.theNumberOfJobs * self.eventsPerJob
|
| 351 |
|
|
|
| 352 |
|
|
# If user requested all the events in the dataset
|
| 353 |
|
|
if (totalEventsRequested == -1):
|
| 354 |
|
|
eventsRemaining=self.maxEvents
|
| 355 |
|
|
# If user requested more events than are in the dataset
|
| 356 |
|
|
elif (totalEventsRequested > self.maxEvents):
|
| 357 |
|
|
eventsRemaining = self.maxEvents
|
| 358 |
|
|
common.logger.message("Requested "+str(self.total_number_of_events)+ " events, but only "+str(self.maxEvents)+" events are available.")
|
| 359 |
|
|
# If user requested less events than are in the dataset
|
| 360 |
|
|
else:
|
| 361 |
|
|
eventsRemaining = totalEventsRequested
|
| 362 |
slacapra |
1.22 |
|
| 363 |
slacapra |
1.41 |
# If user requested more events per job than are in the dataset
|
| 364 |
|
|
if (self.selectEventsPerJob and eventsPerJobRequested > self.maxEvents):
|
| 365 |
|
|
eventsPerJobRequested = self.maxEvents
|
| 366 |
|
|
|
| 367 |
gutsche |
1.35 |
# For user info at end
|
| 368 |
|
|
totalEventCount = 0
|
| 369 |
gutsche |
1.3 |
|
| 370 |
gutsche |
1.35 |
if (self.selectTotalNumberEvents and self.selectNumberOfJobs):
|
| 371 |
|
|
eventsPerJobRequested = int(eventsRemaining/self.theNumberOfJobs)
|
| 372 |
slacapra |
1.22 |
|
| 373 |
gutsche |
1.35 |
if (self.selectNumberOfJobs):
|
| 374 |
|
|
common.logger.message("May not create the exact number_of_jobs requested.")
|
| 375 |
slacapra |
1.23 |
|
| 376 |
gutsche |
1.38 |
if ( self.ncjobs == 'all' ) :
|
| 377 |
|
|
totalNumberOfJobs = 999999999
|
| 378 |
|
|
else :
|
| 379 |
|
|
totalNumberOfJobs = self.ncjobs
|
| 380 |
|
|
|
| 381 |
|
|
|
| 382 |
gutsche |
1.35 |
blocks = blockSites.keys()
|
| 383 |
|
|
blockCount = 0
|
| 384 |
|
|
# Backup variable in case self.maxEvents counted events in a non-included block
|
| 385 |
|
|
numBlocksInDataset = len(blocks)
|
| 386 |
gutsche |
1.3 |
|
| 387 |
gutsche |
1.35 |
jobCount = 0
|
| 388 |
|
|
list_of_lists = []
|
| 389 |
gutsche |
1.3 |
|
| 390 |
gutsche |
1.35 |
# ---- Iterate over the blocks in the dataset until ---- #
|
| 391 |
|
|
# ---- we've met the requested total # of events ---- #
|
| 392 |
gutsche |
1.38 |
while ( (eventsRemaining > 0) and (blockCount < numBlocksInDataset) and (jobCount < totalNumberOfJobs)):
|
| 393 |
gutsche |
1.35 |
block = blocks[blockCount]
|
| 394 |
gutsche |
1.44 |
blockCount += 1
|
| 395 |
|
|
|
| 396 |
gutsche |
1.3 |
|
| 397 |
gutsche |
1.44 |
numEventsInBlock = self.eventsbyblock[block]
|
| 398 |
|
|
common.logger.debug(5,'Events in Block File '+str(numEventsInBlock))
|
| 399 |
slacapra |
1.9 |
|
| 400 |
gutsche |
1.35 |
files = self.filesbyblock[block]
|
| 401 |
|
|
numFilesInBlock = len(files)
|
| 402 |
|
|
if (numFilesInBlock <= 0):
|
| 403 |
|
|
continue
|
| 404 |
|
|
fileCount = 0
|
| 405 |
|
|
|
| 406 |
|
|
# ---- New block => New job ---- #
|
| 407 |
|
|
parString = "\\{"
|
| 408 |
gutsche |
1.38 |
# counter for number of events in files currently worked on
|
| 409 |
|
|
filesEventCount = 0
|
| 410 |
|
|
# flag if next while loop should touch new file
|
| 411 |
|
|
newFile = 1
|
| 412 |
|
|
# job event counter
|
| 413 |
|
|
jobSkipEventCount = 0
|
| 414 |
slacapra |
1.9 |
|
| 415 |
gutsche |
1.35 |
# ---- Iterate over the files in the block until we've met the requested ---- #
|
| 416 |
|
|
# ---- total # of events or we've gone over all the files in this block ---- #
|
| 417 |
gutsche |
1.38 |
while ( (eventsRemaining > 0) and (fileCount < numFilesInBlock) and (jobCount < totalNumberOfJobs) ):
|
| 418 |
gutsche |
1.35 |
file = files[fileCount]
|
| 419 |
gutsche |
1.38 |
if newFile :
|
| 420 |
slacapra |
1.41 |
try:
|
| 421 |
|
|
numEventsInFile = self.eventsbyfile[file]
|
| 422 |
|
|
common.logger.debug(6, "File "+str(file)+" has "+str(numEventsInFile)+" events")
|
| 423 |
|
|
# increase filesEventCount
|
| 424 |
|
|
filesEventCount += numEventsInFile
|
| 425 |
|
|
# Add file to current job
|
| 426 |
|
|
parString += '\\\"' + file + '\\\"\,'
|
| 427 |
|
|
newFile = 0
|
| 428 |
|
|
except KeyError:
|
| 429 |
gutsche |
1.44 |
common.logger.message("File "+str(file)+" has unknown number of events: skipping")
|
| 430 |
slacapra |
1.41 |
|
| 431 |
gutsche |
1.38 |
|
| 432 |
|
|
# if less events in file remain than eventsPerJobRequested
|
| 433 |
|
|
if ( filesEventCount - jobSkipEventCount < eventsPerJobRequested ) :
|
| 434 |
|
|
# if last file in block
|
| 435 |
gutsche |
1.44 |
if ( fileCount == numFilesInBlock-1 ) :
|
| 436 |
gutsche |
1.38 |
# end job using last file, use remaining events in block
|
| 437 |
|
|
# close job and touch new file
|
| 438 |
|
|
fullString = parString[:-2]
|
| 439 |
|
|
fullString += '\\}'
|
| 440 |
|
|
list_of_lists.append([fullString,str(-1),str(jobSkipEventCount)])
|
| 441 |
slacapra |
1.41 |
common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(filesEventCount - jobSkipEventCount)+" events (last file in block).")
|
| 442 |
gutsche |
1.38 |
self.jobDestination.append(blockSites[block])
|
| 443 |
|
|
common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(self.jobDestination[jobCount]))
|
| 444 |
|
|
# reset counter
|
| 445 |
|
|
jobCount = jobCount + 1
|
| 446 |
gutsche |
1.44 |
totalEventCount = totalEventCount + filesEventCount - jobSkipEventCount
|
| 447 |
|
|
eventsRemaining = eventsRemaining - filesEventCount + jobSkipEventCount
|
| 448 |
gutsche |
1.38 |
jobSkipEventCount = 0
|
| 449 |
|
|
# reset file
|
| 450 |
|
|
parString = "\\{"
|
| 451 |
|
|
filesEventCount = 0
|
| 452 |
|
|
newFile = 1
|
| 453 |
|
|
fileCount += 1
|
| 454 |
|
|
else :
|
| 455 |
|
|
# go to next file
|
| 456 |
|
|
newFile = 1
|
| 457 |
|
|
fileCount += 1
|
| 458 |
|
|
# if events in file equal to eventsPerJobRequested
|
| 459 |
|
|
elif ( filesEventCount - jobSkipEventCount == eventsPerJobRequested ) :
|
| 460 |
|
|
# close job and touch new file
|
| 461 |
gutsche |
1.35 |
fullString = parString[:-2]
|
| 462 |
|
|
fullString += '\\}'
|
| 463 |
gutsche |
1.38 |
list_of_lists.append([fullString,str(eventsPerJobRequested),str(jobSkipEventCount)])
|
| 464 |
slacapra |
1.41 |
common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(eventsPerJobRequested)+" events.")
|
| 465 |
gutsche |
1.38 |
self.jobDestination.append(blockSites[block])
|
| 466 |
|
|
common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(self.jobDestination[jobCount]))
|
| 467 |
|
|
# reset counter
|
| 468 |
|
|
jobCount = jobCount + 1
|
| 469 |
|
|
totalEventCount = totalEventCount + eventsPerJobRequested
|
| 470 |
|
|
eventsRemaining = eventsRemaining - eventsPerJobRequested
|
| 471 |
|
|
jobSkipEventCount = 0
|
| 472 |
|
|
# reset file
|
| 473 |
|
|
parString = "\\{"
|
| 474 |
|
|
filesEventCount = 0
|
| 475 |
|
|
newFile = 1
|
| 476 |
|
|
fileCount += 1
|
| 477 |
|
|
|
| 478 |
|
|
# if more events in file remain than eventsPerJobRequested
|
| 479 |
|
|
else :
|
| 480 |
|
|
# close job but don't touch new file
|
| 481 |
|
|
fullString = parString[:-2]
|
| 482 |
|
|
fullString += '\\}'
|
| 483 |
|
|
list_of_lists.append([fullString,str(eventsPerJobRequested),str(jobSkipEventCount)])
|
| 484 |
slacapra |
1.41 |
common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(eventsPerJobRequested)+" events.")
|
| 485 |
gutsche |
1.35 |
self.jobDestination.append(blockSites[block])
|
| 486 |
|
|
common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(self.jobDestination[jobCount]))
|
| 487 |
gutsche |
1.38 |
# increase counter
|
| 488 |
|
|
jobCount = jobCount + 1
|
| 489 |
|
|
totalEventCount = totalEventCount + eventsPerJobRequested
|
| 490 |
|
|
eventsRemaining = eventsRemaining - eventsPerJobRequested
|
| 491 |
|
|
# calculate skip events for last file
|
| 492 |
|
|
# use filesEventCount (contains several files), jobSkipEventCount and eventsPerJobRequest
|
| 493 |
|
|
jobSkipEventCount = eventsPerJobRequested - (filesEventCount - jobSkipEventCount - self.eventsbyfile[file])
|
| 494 |
|
|
# remove all but the last file
|
| 495 |
|
|
filesEventCount = self.eventsbyfile[file]
|
| 496 |
|
|
parString = "\\{"
|
| 497 |
|
|
parString += '\\\"' + file + '\\\"\,'
|
| 498 |
slacapra |
1.41 |
pass # END if
|
| 499 |
gutsche |
1.35 |
pass # END while (iterate over files in the block)
|
| 500 |
|
|
pass # END while (iterate over blocks in the dataset)
|
| 501 |
slacapra |
1.41 |
self.ncjobs = self.total_number_of_jobs = jobCount
|
| 502 |
gutsche |
1.38 |
if (eventsRemaining > 0 and jobCount < totalNumberOfJobs ):
|
| 503 |
gutsche |
1.35 |
common.logger.message("Could not run on all requested events because some blocks not hosted at allowed sites.")
|
| 504 |
mkirn |
1.37 |
common.logger.message("\n"+str(jobCount)+" job(s) can run on "+str(totalEventCount)+" events.\n")
|
| 505 |
slacapra |
1.22 |
|
| 506 |
slacapra |
1.9 |
self.list_of_args = list_of_lists
|
| 507 |
|
|
return
|
| 508 |
|
|
|
| 509 |
slacapra |
1.21 |
def jobSplittingNoInput(self):
|
| 510 |
slacapra |
1.9 |
"""
|
| 511 |
|
|
Perform job splitting based on number of event per job
|
| 512 |
|
|
"""
|
| 513 |
|
|
common.logger.debug(5,'Splitting per events')
|
| 514 |
|
|
common.logger.message('Required '+str(self.eventsPerJob)+' events per job ')
|
| 515 |
slacapra |
1.22 |
common.logger.message('Required '+str(self.theNumberOfJobs)+' jobs in total ')
|
| 516 |
slacapra |
1.9 |
common.logger.message('Required '+str(self.total_number_of_events)+' events in total ')
|
| 517 |
|
|
|
| 518 |
slacapra |
1.10 |
if (self.total_number_of_events < 0):
|
| 519 |
|
|
msg='Cannot split jobs per Events with "-1" as total number of events'
|
| 520 |
|
|
raise CrabException(msg)
|
| 521 |
|
|
|
| 522 |
slacapra |
1.22 |
if (self.selectEventsPerJob):
|
| 523 |
spiga |
1.65 |
if (self.selectTotalNumberEvents):
|
| 524 |
|
|
self.total_number_of_jobs = int(self.total_number_of_events/self.eventsPerJob)
|
| 525 |
|
|
elif(self.selectNumberOfJobs) :
|
| 526 |
|
|
self.total_number_of_jobs =self.theNumberOfJobs
|
| 527 |
|
|
self.total_number_of_events =int(self.theNumberOfJobs*self.eventsPerJob)
|
| 528 |
|
|
|
| 529 |
slacapra |
1.22 |
elif (self.selectNumberOfJobs) :
|
| 530 |
|
|
self.total_number_of_jobs = self.theNumberOfJobs
|
| 531 |
|
|
self.eventsPerJob = int(self.total_number_of_events/self.total_number_of_jobs)
|
| 532 |
spiga |
1.65 |
|
| 533 |
slacapra |
1.9 |
common.logger.debug(5,'N jobs '+str(self.total_number_of_jobs))
|
| 534 |
|
|
|
| 535 |
|
|
# is there any remainder?
|
| 536 |
|
|
check = int(self.total_number_of_events) - (int(self.total_number_of_jobs)*self.eventsPerJob)
|
| 537 |
|
|
|
| 538 |
|
|
common.logger.debug(5,'Check '+str(check))
|
| 539 |
|
|
|
| 540 |
gutsche |
1.35 |
common.logger.message(str(self.total_number_of_jobs)+' jobs can be created, each for '+str(self.eventsPerJob)+' for a total of '+str(self.total_number_of_jobs*self.eventsPerJob)+' events')
|
| 541 |
slacapra |
1.9 |
if check > 0:
|
| 542 |
gutsche |
1.35 |
common.logger.message('Warning: asked '+str(self.total_number_of_events)+' but can do only '+str(int(self.total_number_of_jobs)*self.eventsPerJob))
|
| 543 |
slacapra |
1.9 |
|
| 544 |
slacapra |
1.10 |
# argument is seed number.$i
|
| 545 |
slacapra |
1.9 |
self.list_of_args = []
|
| 546 |
|
|
for i in range(self.total_number_of_jobs):
|
| 547 |
gutsche |
1.35 |
## Since there is no input, any site is good
|
| 548 |
spiga |
1.42 |
# self.jobDestination.append(["Any"])
|
| 549 |
|
|
self.jobDestination.append([""]) #must be empty to write correctly the xml
|
| 550 |
spiga |
1.57 |
args=''
|
| 551 |
|
|
if (self.firstRun):
|
| 552 |
|
|
## pythia first run
|
| 553 |
|
|
#self.list_of_args.append([(str(self.firstRun)+str(i))])
|
| 554 |
|
|
args=args+(str(self.firstRun)+str(i))
|
| 555 |
|
|
else:
|
| 556 |
|
|
## no first run
|
| 557 |
|
|
#self.list_of_args.append([str(i)])
|
| 558 |
|
|
args=args+str(i)
|
| 559 |
slacapra |
1.23 |
if (self.sourceSeed):
|
| 560 |
slacapra |
1.28 |
if (self.sourceSeedVtx):
|
| 561 |
|
|
## pythia + vtx random seed
|
| 562 |
spiga |
1.57 |
#self.list_of_args.append([
|
| 563 |
|
|
# str(self.sourceSeed)+str(i),
|
| 564 |
|
|
# str(self.sourceSeedVtx)+str(i)
|
| 565 |
|
|
# ])
|
| 566 |
|
|
args=args+str(',')+str(self.sourceSeed)+str(i)+str(',')+str(self.sourceSeedVtx)+str(i)
|
| 567 |
slacapra |
1.28 |
else:
|
| 568 |
|
|
## only pythia random seed
|
| 569 |
spiga |
1.57 |
#self.list_of_args.append([(str(self.sourceSeed)+str(i))])
|
| 570 |
|
|
args=args +str(',')+str(self.sourceSeed)+str(i)
|
| 571 |
slacapra |
1.23 |
else:
|
| 572 |
slacapra |
1.28 |
## no random seed
|
| 573 |
spiga |
1.57 |
if str(args)=='': args=args+(str(self.firstRun)+str(i))
|
| 574 |
|
|
arguments=args.split(',')
|
| 575 |
|
|
if len(arguments)==3:self.list_of_args.append([str(arguments[0]),str(arguments[1]),str(arguments[2])])
|
| 576 |
|
|
elif len(arguments)==2:self.list_of_args.append([str(arguments[0]),str(arguments[1])])
|
| 577 |
|
|
else :self.list_of_args.append([str(arguments[0])])
|
| 578 |
|
|
|
| 579 |
|
|
# print self.list_of_args
|
| 580 |
gutsche |
1.3 |
|
| 581 |
|
|
return
|
| 582 |
|
|
|
| 583 |
spiga |
1.42 |
|
| 584 |
|
|
def jobSplittingForScript(self):#CarlosDaniele
|
| 585 |
|
|
"""
|
| 586 |
|
|
Perform job splitting based on number of job
|
| 587 |
|
|
"""
|
| 588 |
|
|
common.logger.debug(5,'Splitting per job')
|
| 589 |
|
|
common.logger.message('Required '+str(self.theNumberOfJobs)+' jobs in total ')
|
| 590 |
|
|
|
| 591 |
|
|
self.total_number_of_jobs = self.theNumberOfJobs
|
| 592 |
|
|
|
| 593 |
|
|
common.logger.debug(5,'N jobs '+str(self.total_number_of_jobs))
|
| 594 |
|
|
|
| 595 |
|
|
common.logger.message(str(self.total_number_of_jobs)+' jobs can be created')
|
| 596 |
|
|
|
| 597 |
|
|
# argument is seed number.$i
|
| 598 |
|
|
self.list_of_args = []
|
| 599 |
|
|
for i in range(self.total_number_of_jobs):
|
| 600 |
|
|
## Since there is no input, any site is good
|
| 601 |
|
|
# self.jobDestination.append(["Any"])
|
| 602 |
|
|
self.jobDestination.append([""])
|
| 603 |
|
|
## no random seed
|
| 604 |
|
|
self.list_of_args.append([str(i)])
|
| 605 |
|
|
return
|
| 606 |
|
|
|
| 607 |
gutsche |
1.3 |
def split(self, jobParams):
|
| 608 |
|
|
|
| 609 |
|
|
common.jobDB.load()
|
| 610 |
|
|
#### Fabio
|
| 611 |
|
|
njobs = self.total_number_of_jobs
|
| 612 |
slacapra |
1.9 |
arglist = self.list_of_args
|
| 613 |
gutsche |
1.3 |
# create the empty structure
|
| 614 |
|
|
for i in range(njobs):
|
| 615 |
|
|
jobParams.append("")
|
| 616 |
|
|
|
| 617 |
|
|
for job in range(njobs):
|
| 618 |
slacapra |
1.17 |
jobParams[job] = arglist[job]
|
| 619 |
|
|
# print str(arglist[job])
|
| 620 |
|
|
# print jobParams[job]
|
| 621 |
gutsche |
1.3 |
common.jobDB.setArguments(job, jobParams[job])
|
| 622 |
gutsche |
1.35 |
common.logger.debug(5,"Job "+str(job)+" Destination: "+str(self.jobDestination[job]))
|
| 623 |
|
|
common.jobDB.setDestination(job, self.jobDestination[job])
|
| 624 |
gutsche |
1.3 |
|
| 625 |
|
|
common.jobDB.save()
|
| 626 |
|
|
return
|
| 627 |
|
|
|
| 628 |
|
|
def getJobTypeArguments(self, nj, sched):
|
| 629 |
slacapra |
1.17 |
result = ''
|
| 630 |
|
|
for i in common.jobDB.arguments(nj):
|
| 631 |
|
|
result=result+str(i)+" "
|
| 632 |
|
|
return result
|
| 633 |
gutsche |
1.3 |
|
| 634 |
|
|
def numberOfJobs(self):
|
| 635 |
|
|
# Fabio
|
| 636 |
|
|
return self.total_number_of_jobs
|
| 637 |
|
|
|
| 638 |
slacapra |
1.1 |
def getTarBall(self, exe):
|
| 639 |
|
|
"""
|
| 640 |
|
|
Return the TarBall with lib and exe
|
| 641 |
|
|
"""
|
| 642 |
|
|
|
| 643 |
|
|
# if it exist, just return it
|
| 644 |
corvo |
1.56 |
#
|
| 645 |
|
|
# Marco. Let's start to use relative path for Boss XML files
|
| 646 |
|
|
#
|
| 647 |
|
|
self.tgzNameWithPath = common.work_space.pathForTgz()+'share/'+self.tgz_name
|
| 648 |
slacapra |
1.1 |
if os.path.exists(self.tgzNameWithPath):
|
| 649 |
|
|
return self.tgzNameWithPath
|
| 650 |
|
|
|
| 651 |
|
|
# Prepare a tar gzipped file with user binaries.
|
| 652 |
|
|
self.buildTar_(exe)
|
| 653 |
|
|
|
| 654 |
|
|
return string.strip(self.tgzNameWithPath)
|
| 655 |
|
|
|
| 656 |
|
|
def buildTar_(self, executable):
|
| 657 |
|
|
|
| 658 |
|
|
# First of all declare the user Scram area
|
| 659 |
|
|
swArea = self.scram.getSWArea_()
|
| 660 |
|
|
#print "swArea = ", swArea
|
| 661 |
slacapra |
1.63 |
# swVersion = self.scram.getSWVersion()
|
| 662 |
|
|
# print "swVersion = ", swVersion
|
| 663 |
slacapra |
1.1 |
swReleaseTop = self.scram.getReleaseTop_()
|
| 664 |
|
|
#print "swReleaseTop = ", swReleaseTop
|
| 665 |
|
|
|
| 666 |
|
|
## check if working area is release top
|
| 667 |
|
|
if swReleaseTop == '' or swArea == swReleaseTop:
|
| 668 |
|
|
return
|
| 669 |
|
|
|
| 670 |
slacapra |
1.61 |
import tarfile
|
| 671 |
|
|
try: # create tar ball
|
| 672 |
|
|
tar = tarfile.open(self.tgzNameWithPath, "w:gz")
|
| 673 |
|
|
## First find the executable
|
| 674 |
|
|
if (self.executable != ''):
|
| 675 |
|
|
exeWithPath = self.scram.findFile_(executable)
|
| 676 |
|
|
if ( not exeWithPath ):
|
| 677 |
|
|
raise CrabException('User executable '+executable+' not found')
|
| 678 |
|
|
|
| 679 |
|
|
## then check if it's private or not
|
| 680 |
|
|
if exeWithPath.find(swReleaseTop) == -1:
|
| 681 |
|
|
# the exe is private, so we must ship
|
| 682 |
|
|
common.logger.debug(5,"Exe "+exeWithPath+" to be tarred")
|
| 683 |
|
|
path = swArea+'/'
|
| 684 |
|
|
exe = string.replace(exeWithPath, path,'')
|
| 685 |
|
|
tar.add(path+exe,exe)
|
| 686 |
|
|
pass
|
| 687 |
|
|
else:
|
| 688 |
|
|
# the exe is from release, we'll find it on WN
|
| 689 |
|
|
pass
|
| 690 |
|
|
|
| 691 |
|
|
## Now get the libraries: only those in local working area
|
| 692 |
|
|
libDir = 'lib'
|
| 693 |
|
|
lib = swArea+'/' +libDir
|
| 694 |
|
|
common.logger.debug(5,"lib "+lib+" to be tarred")
|
| 695 |
|
|
if os.path.exists(lib):
|
| 696 |
|
|
tar.add(lib,libDir)
|
| 697 |
|
|
|
| 698 |
|
|
## Now check if module dir is present
|
| 699 |
|
|
moduleDir = 'module'
|
| 700 |
|
|
module = swArea + '/' + moduleDir
|
| 701 |
|
|
if os.path.isdir(module):
|
| 702 |
|
|
tar.add(module,moduleDir)
|
| 703 |
|
|
|
| 704 |
|
|
## Now check if any data dir(s) is present
|
| 705 |
|
|
swAreaLen=len(swArea)
|
| 706 |
|
|
for root, dirs, files in os.walk(swArea):
|
| 707 |
|
|
if "data" in dirs:
|
| 708 |
|
|
common.logger.debug(5,"data "+root+"/data"+" to be tarred")
|
| 709 |
|
|
tar.add(root+"/data",root[swAreaLen:]+"/data")
|
| 710 |
|
|
|
| 711 |
|
|
## Add ProdAgent dir to tar
|
| 712 |
|
|
paDir = 'ProdAgentApi'
|
| 713 |
|
|
pa = os.environ['CRABDIR'] + '/' + 'ProdAgentApi'
|
| 714 |
|
|
if os.path.isdir(pa):
|
| 715 |
|
|
tar.add(pa,paDir)
|
| 716 |
|
|
|
| 717 |
|
|
common.logger.debug(5,"Files added to "+self.tgzNameWithPath+" : "+str(tar.getnames()))
|
| 718 |
|
|
tar.close()
|
| 719 |
|
|
except :
|
| 720 |
|
|
raise CrabException('Could not create tar-ball')
|
| 721 |
corvo |
1.56 |
|
| 722 |
slacapra |
1.61 |
## create tar-ball with ML stuff
|
| 723 |
corvo |
1.58 |
self.MLtgzfile = common.work_space.pathForTgz()+'share/MLfiles.tgz'
|
| 724 |
slacapra |
1.61 |
try:
|
| 725 |
|
|
tar = tarfile.open(self.MLtgzfile, "w:gz")
|
| 726 |
|
|
path=os.environ['CRABDIR'] + '/python/'
|
| 727 |
|
|
for file in ['report.py', 'DashboardAPI.py', 'Logger.py', 'ProcInfo.py', 'apmon.py', 'parseCrabFjr.py']:
|
| 728 |
|
|
tar.add(path+file,file)
|
| 729 |
|
|
common.logger.debug(5,"Files added to "+self.MLtgzfile+" : "+str(tar.getnames()))
|
| 730 |
|
|
tar.close()
|
| 731 |
|
|
except :
|
| 732 |
corvo |
1.58 |
raise CrabException('Could not create ML files tar-ball')
|
| 733 |
|
|
|
| 734 |
slacapra |
1.1 |
return
|
| 735 |
|
|
|
| 736 |
|
|
def wsSetupEnvironment(self, nj):
|
| 737 |
|
|
"""
|
| 738 |
|
|
Returns part of a job script which prepares
|
| 739 |
|
|
the execution environment for the job 'nj'.
|
| 740 |
|
|
"""
|
| 741 |
|
|
# Prepare JobType-independent part
|
| 742 |
gutsche |
1.3 |
txt = ''
|
| 743 |
|
|
|
| 744 |
|
|
## OLI_Daniele at this level middleware already known
|
| 745 |
|
|
|
| 746 |
|
|
txt += 'if [ $middleware == LCG ]; then \n'
|
| 747 |
|
|
txt += self.wsSetupCMSLCGEnvironment_()
|
| 748 |
|
|
txt += 'elif [ $middleware == OSG ]; then\n'
|
| 749 |
gutsche |
1.43 |
txt += ' WORKING_DIR=`/bin/mktemp -d $OSG_WN_TMP/cms_XXXXXXXXXXXX`\n'
|
| 750 |
|
|
txt += ' echo "Created working directory: $WORKING_DIR"\n'
|
| 751 |
gutsche |
1.3 |
txt += ' if [ ! -d $WORKING_DIR ] ;then\n'
|
| 752 |
gutsche |
1.7 |
txt += ' echo "SET_CMS_ENV 10016 ==> OSG $WORKING_DIR could not be created on WN `hostname`"\n'
|
| 753 |
|
|
txt += ' echo "JOB_EXIT_STATUS = 10016"\n'
|
| 754 |
|
|
txt += ' echo "JobExitCode=10016" | tee -a $RUNTIME_AREA/$repo\n'
|
| 755 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 756 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 757 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 758 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 759 |
gutsche |
1.3 |
txt += ' exit 1\n'
|
| 760 |
|
|
txt += ' fi\n'
|
| 761 |
|
|
txt += '\n'
|
| 762 |
|
|
txt += ' echo "Change to working directory: $WORKING_DIR"\n'
|
| 763 |
|
|
txt += ' cd $WORKING_DIR\n'
|
| 764 |
|
|
txt += self.wsSetupCMSOSGEnvironment_()
|
| 765 |
|
|
txt += 'fi\n'
|
| 766 |
slacapra |
1.1 |
|
| 767 |
|
|
# Prepare JobType-specific part
|
| 768 |
|
|
scram = self.scram.commandName()
|
| 769 |
|
|
txt += '\n\n'
|
| 770 |
|
|
txt += 'echo "### SPECIFIC JOB SETUP ENVIRONMENT ###"\n'
|
| 771 |
|
|
txt += scram+' project CMSSW '+self.version+'\n'
|
| 772 |
|
|
txt += 'status=$?\n'
|
| 773 |
|
|
txt += 'if [ $status != 0 ] ; then\n'
|
| 774 |
gutsche |
1.7 |
txt += ' echo "SET_EXE_ENV 10034 ==>ERROR CMSSW '+self.version+' not found on `hostname`" \n'
|
| 775 |
gutsche |
1.3 |
txt += ' echo "JOB_EXIT_STATUS = 10034"\n'
|
| 776 |
gutsche |
1.7 |
txt += ' echo "JobExitCode=10034" | tee -a $RUNTIME_AREA/$repo\n'
|
| 777 |
slacapra |
1.1 |
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 778 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 779 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 780 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 781 |
gutsche |
1.3 |
## OLI_Daniele
|
| 782 |
|
|
txt += ' if [ $middleware == OSG ]; then \n'
|
| 783 |
|
|
txt += ' echo "Remove working directory: $WORKING_DIR"\n'
|
| 784 |
|
|
txt += ' cd $RUNTIME_AREA\n'
|
| 785 |
|
|
txt += ' /bin/rm -rf $WORKING_DIR\n'
|
| 786 |
|
|
txt += ' if [ -d $WORKING_DIR ] ;then\n'
|
| 787 |
gutsche |
1.7 |
txt += ' echo "SET_CMS_ENV 10018 ==> OSG $WORKING_DIR could not be deleted on WN `hostname` after CMSSW CMSSW_0_6_1 not found on `hostname`"\n'
|
| 788 |
|
|
txt += ' echo "JOB_EXIT_STATUS = 10018"\n'
|
| 789 |
|
|
txt += ' echo "JobExitCode=10018" | tee -a $RUNTIME_AREA/$repo\n'
|
| 790 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 791 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 792 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 793 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 794 |
gutsche |
1.3 |
txt += ' fi\n'
|
| 795 |
|
|
txt += ' fi \n'
|
| 796 |
|
|
txt += ' exit 1 \n'
|
| 797 |
slacapra |
1.1 |
txt += 'fi \n'
|
| 798 |
|
|
txt += 'echo "CMSSW_VERSION = '+self.version+'"\n'
|
| 799 |
|
|
txt += 'cd '+self.version+'\n'
|
| 800 |
|
|
### needed grep for bug in scramv1 ###
|
| 801 |
corvo |
1.58 |
txt += scram+' runtime -sh\n'
|
| 802 |
slacapra |
1.1 |
txt += 'eval `'+scram+' runtime -sh | grep -v SCRAMRT_LSB_JOBNAME`\n'
|
| 803 |
corvo |
1.58 |
txt += 'echo $PATH\n'
|
| 804 |
slacapra |
1.1 |
|
| 805 |
|
|
# Handle the arguments:
|
| 806 |
|
|
txt += "\n"
|
| 807 |
gutsche |
1.7 |
txt += "## number of arguments (first argument always jobnumber)\n"
|
| 808 |
slacapra |
1.1 |
txt += "\n"
|
| 809 |
mkirn |
1.32 |
# txt += "narg=$#\n"
|
| 810 |
|
|
txt += "if [ $nargs -lt 2 ]\n"
|
| 811 |
slacapra |
1.1 |
txt += "then\n"
|
| 812 |
mkirn |
1.33 |
txt += " echo 'SET_EXE_ENV 1 ==> ERROR Too few arguments' +$nargs+ \n"
|
| 813 |
gutsche |
1.3 |
txt += ' echo "JOB_EXIT_STATUS = 50113"\n'
|
| 814 |
gutsche |
1.7 |
txt += ' echo "JobExitCode=50113" | tee -a $RUNTIME_AREA/$repo\n'
|
| 815 |
slacapra |
1.1 |
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 816 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 817 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 818 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 819 |
gutsche |
1.3 |
## OLI_Daniele
|
| 820 |
|
|
txt += ' if [ $middleware == OSG ]; then \n'
|
| 821 |
|
|
txt += ' echo "Remove working directory: $WORKING_DIR"\n'
|
| 822 |
|
|
txt += ' cd $RUNTIME_AREA\n'
|
| 823 |
|
|
txt += ' /bin/rm -rf $WORKING_DIR\n'
|
| 824 |
|
|
txt += ' if [ -d $WORKING_DIR ] ;then\n'
|
| 825 |
gutsche |
1.7 |
txt += ' echo "SET_EXE_ENV 50114 ==> OSG $WORKING_DIR could not be deleted on WN `hostname` after Too few arguments for CRAB job wrapper"\n'
|
| 826 |
|
|
txt += ' echo "JOB_EXIT_STATUS = 50114"\n'
|
| 827 |
|
|
txt += ' echo "JobExitCode=50114" | tee -a $RUNTIME_AREA/$repo\n'
|
| 828 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 829 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 830 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 831 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 832 |
gutsche |
1.3 |
txt += ' fi\n'
|
| 833 |
|
|
txt += ' fi \n'
|
| 834 |
slacapra |
1.1 |
txt += " exit 1\n"
|
| 835 |
|
|
txt += "fi\n"
|
| 836 |
|
|
txt += "\n"
|
| 837 |
|
|
|
| 838 |
|
|
# Prepare job-specific part
|
| 839 |
|
|
job = common.job_list[nj]
|
| 840 |
spiga |
1.42 |
if self.pset != None: #CarlosDaniele
|
| 841 |
|
|
pset = os.path.basename(job.configFilename())
|
| 842 |
|
|
txt += '\n'
|
| 843 |
|
|
if (self.datasetPath): # standard job
|
| 844 |
|
|
#txt += 'InputFiles=$2\n'
|
| 845 |
|
|
txt += 'InputFiles=${args[1]}\n'
|
| 846 |
|
|
txt += 'MaxEvents=${args[2]}\n'
|
| 847 |
|
|
txt += 'SkipEvents=${args[3]}\n'
|
| 848 |
|
|
txt += 'echo "Inputfiles:<$InputFiles>"\n'
|
| 849 |
|
|
txt += 'sed "s#{\'INPUT\'}#$InputFiles#" $RUNTIME_AREA/'+pset+' > pset_tmp_1.cfg\n'
|
| 850 |
|
|
txt += 'echo "MaxEvents:<$MaxEvents>"\n'
|
| 851 |
gutsche |
1.46 |
txt += 'sed "s#INPUTMAXEVENTS#$MaxEvents#" pset_tmp_1.cfg > pset_tmp_2.cfg\n'
|
| 852 |
spiga |
1.42 |
txt += 'echo "SkipEvents:<$SkipEvents>"\n'
|
| 853 |
gutsche |
1.46 |
txt += 'sed "s#INPUTSKIPEVENTS#$SkipEvents#" pset_tmp_2.cfg > pset.cfg\n'
|
| 854 |
spiga |
1.42 |
else: # pythia like job
|
| 855 |
|
|
if (self.sourceSeed):
|
| 856 |
spiga |
1.57 |
txt += 'FirstRun=${args[1]}\n'
|
| 857 |
|
|
txt += 'echo "FirstRun: <$FirstRun>"\n'
|
| 858 |
|
|
txt += 'sed "s#\<INPUTFIRSTRUN\>#$FirstRun#" $RUNTIME_AREA/'+pset+' > tmp_1.cfg\n'
|
| 859 |
|
|
else:
|
| 860 |
|
|
txt += '# Copy untouched pset\n'
|
| 861 |
|
|
txt += 'cp $RUNTIME_AREA/'+pset+' tmp_1.cfg\n'
|
| 862 |
|
|
if (self.sourceSeed):
|
| 863 |
spiga |
1.42 |
# txt += 'Seed=$2\n'
|
| 864 |
spiga |
1.57 |
txt += 'Seed=${args[2]}\n'
|
| 865 |
spiga |
1.42 |
txt += 'echo "Seed: <$Seed>"\n'
|
| 866 |
spiga |
1.57 |
txt += 'sed "s#\<INPUT\>#$Seed#" tmp_1.cfg > tmp_2.cfg\n'
|
| 867 |
spiga |
1.42 |
if (self.sourceSeedVtx):
|
| 868 |
|
|
# txt += 'VtxSeed=$3\n'
|
| 869 |
spiga |
1.57 |
txt += 'VtxSeed=${args[3]}\n'
|
| 870 |
spiga |
1.42 |
txt += 'echo "VtxSeed: <$VtxSeed>"\n'
|
| 871 |
spiga |
1.57 |
txt += 'sed "s#INPUTVTX#$VtxSeed#" tmp_2.cfg > pset.cfg\n'
|
| 872 |
spiga |
1.42 |
else:
|
| 873 |
spiga |
1.57 |
txt += 'mv tmp_2.cfg pset.cfg\n'
|
| 874 |
slacapra |
1.28 |
else:
|
| 875 |
spiga |
1.57 |
txt += 'mv tmp_1.cfg pset.cfg\n'
|
| 876 |
|
|
# txt += '# Copy untouched pset\n'
|
| 877 |
|
|
# txt += 'cp $RUNTIME_AREA/'+pset+' pset.cfg\n'
|
| 878 |
slacapra |
1.24 |
|
| 879 |
slacapra |
1.1 |
|
| 880 |
|
|
if len(self.additional_inbox_files) > 0:
|
| 881 |
|
|
for file in self.additional_inbox_files:
|
| 882 |
mkirn |
1.31 |
relFile = file.split("/")[-1]
|
| 883 |
|
|
txt += 'if [ -e $RUNTIME_AREA/'+relFile+' ] ; then\n'
|
| 884 |
|
|
txt += ' cp $RUNTIME_AREA/'+relFile+' .\n'
|
| 885 |
|
|
txt += ' chmod +x '+relFile+'\n'
|
| 886 |
slacapra |
1.1 |
txt += 'fi\n'
|
| 887 |
|
|
pass
|
| 888 |
|
|
|
| 889 |
spiga |
1.42 |
if self.pset != None: #CarlosDaniele
|
| 890 |
|
|
txt += 'echo "### END JOB SETUP ENVIRONMENT ###"\n\n'
|
| 891 |
|
|
|
| 892 |
|
|
txt += '\n'
|
| 893 |
|
|
txt += 'echo "***** cat pset.cfg *********"\n'
|
| 894 |
|
|
txt += 'cat pset.cfg\n'
|
| 895 |
|
|
txt += 'echo "****** end pset.cfg ********"\n'
|
| 896 |
|
|
txt += '\n'
|
| 897 |
|
|
# txt += 'echo "***** cat pset1.cfg *********"\n'
|
| 898 |
|
|
# txt += 'cat pset1.cfg\n'
|
| 899 |
|
|
# txt += 'echo "****** end pset1.cfg ********"\n'
|
| 900 |
gutsche |
1.3 |
return txt
|
| 901 |
|
|
|
| 902 |
slacapra |
1.63 |
def wsBuildExe(self, nj=0):
|
| 903 |
gutsche |
1.3 |
"""
|
| 904 |
|
|
Put in the script the commands to build an executable
|
| 905 |
|
|
or a library.
|
| 906 |
|
|
"""
|
| 907 |
|
|
|
| 908 |
|
|
txt = ""
|
| 909 |
|
|
|
| 910 |
|
|
if os.path.isfile(self.tgzNameWithPath):
|
| 911 |
|
|
txt += 'echo "tar xzvf $RUNTIME_AREA/'+os.path.basename(self.tgzNameWithPath)+'"\n'
|
| 912 |
|
|
txt += 'tar xzvf $RUNTIME_AREA/'+os.path.basename(self.tgzNameWithPath)+'\n'
|
| 913 |
|
|
txt += 'untar_status=$? \n'
|
| 914 |
|
|
txt += 'if [ $untar_status -ne 0 ]; then \n'
|
| 915 |
|
|
txt += ' echo "SET_EXE 1 ==> ERROR Untarring .tgz file failed"\n'
|
| 916 |
|
|
txt += ' echo "JOB_EXIT_STATUS = $untar_status" \n'
|
| 917 |
gutsche |
1.7 |
txt += ' echo "JobExitCode=$untar_status" | tee -a $RUNTIME_AREA/$repo\n'
|
| 918 |
gutsche |
1.3 |
txt += ' if [ $middleware == OSG ]; then \n'
|
| 919 |
|
|
txt += ' echo "Remove working directory: $WORKING_DIR"\n'
|
| 920 |
|
|
txt += ' cd $RUNTIME_AREA\n'
|
| 921 |
|
|
txt += ' /bin/rm -rf $WORKING_DIR\n'
|
| 922 |
|
|
txt += ' if [ -d $WORKING_DIR ] ;then\n'
|
| 923 |
gutsche |
1.13 |
txt += ' echo "SET_EXE 50999 ==> OSG $WORKING_DIR could not be deleted on WN `hostname` after Untarring .tgz file failed"\n'
|
| 924 |
|
|
txt += ' echo "JOB_EXIT_STATUS = 50999"\n'
|
| 925 |
|
|
txt += ' echo "JobExitCode=50999" | tee -a $RUNTIME_AREA/$repo\n'
|
| 926 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 927 |
|
|
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 928 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 929 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 930 |
gutsche |
1.3 |
txt += ' fi\n'
|
| 931 |
|
|
txt += ' fi \n'
|
| 932 |
|
|
txt += ' \n'
|
| 933 |
gutsche |
1.7 |
txt += ' exit 1 \n'
|
| 934 |
gutsche |
1.3 |
txt += 'else \n'
|
| 935 |
|
|
txt += ' echo "Successful untar" \n'
|
| 936 |
|
|
txt += 'fi \n'
|
| 937 |
gutsche |
1.50 |
txt += '\n'
|
| 938 |
|
|
txt += 'echo "Include ProdAgentApi in PYTHONPATH"\n'
|
| 939 |
|
|
txt += 'if [ -z "$PYTHONPATH" ]; then\n'
|
| 940 |
|
|
txt += ' export PYTHONPATH=ProdAgentApi\n'
|
| 941 |
|
|
txt += 'else\n'
|
| 942 |
|
|
txt += ' export PYTHONPATH=ProdAgentApi:${PYTHONPATH}\n'
|
| 943 |
|
|
txt += 'fi\n'
|
| 944 |
|
|
txt += '\n'
|
| 945 |
|
|
|
| 946 |
gutsche |
1.3 |
pass
|
| 947 |
|
|
|
| 948 |
slacapra |
1.1 |
return txt
|
| 949 |
|
|
|
| 950 |
|
|
def modifySteeringCards(self, nj):
|
| 951 |
|
|
"""
|
| 952 |
|
|
modify the card provided by the user,
|
| 953 |
|
|
writing a new card into share dir
|
| 954 |
|
|
"""
|
| 955 |
|
|
|
| 956 |
|
|
def executableName(self):
|
| 957 |
spiga |
1.42 |
if self.pset == None: #CarlosDaniele
|
| 958 |
|
|
return "sh "
|
| 959 |
|
|
else:
|
| 960 |
|
|
return self.executable
|
| 961 |
slacapra |
1.1 |
|
| 962 |
|
|
def executableArgs(self):
|
| 963 |
spiga |
1.42 |
if self.pset == None:#CarlosDaniele
|
| 964 |
|
|
return self.scriptExe + " $NJob"
|
| 965 |
|
|
else:
|
| 966 |
|
|
return " -p pset.cfg"
|
| 967 |
slacapra |
1.1 |
|
| 968 |
|
|
def inputSandbox(self, nj):
|
| 969 |
|
|
"""
|
| 970 |
|
|
Returns a list of filenames to be put in JDL input sandbox.
|
| 971 |
|
|
"""
|
| 972 |
|
|
inp_box = []
|
| 973 |
slacapra |
1.53 |
# # dict added to delete duplicate from input sandbox file list
|
| 974 |
|
|
# seen = {}
|
| 975 |
slacapra |
1.1 |
## code
|
| 976 |
|
|
if os.path.isfile(self.tgzNameWithPath):
|
| 977 |
|
|
inp_box.append(self.tgzNameWithPath)
|
| 978 |
corvo |
1.58 |
if os.path.isfile(self.MLtgzfile):
|
| 979 |
|
|
inp_box.append(self.MLtgzfile)
|
| 980 |
slacapra |
1.1 |
## config
|
| 981 |
spiga |
1.42 |
if not self.pset is None: #CarlosDaniele
|
| 982 |
corvo |
1.56 |
inp_box.append(common.work_space.pathForTgz() + 'job/' + self.configFilename())
|
| 983 |
slacapra |
1.1 |
## additional input files
|
| 984 |
gutsche |
1.3 |
#for file in self.additional_inbox_files:
|
| 985 |
|
|
# inp_box.append(common.work_space.cwdDir()+file)
|
| 986 |
slacapra |
1.1 |
return inp_box
|
| 987 |
|
|
|
| 988 |
|
|
def outputSandbox(self, nj):
|
| 989 |
|
|
"""
|
| 990 |
|
|
Returns a list of filenames to be put in JDL output sandbox.
|
| 991 |
|
|
"""
|
| 992 |
|
|
out_box = []
|
| 993 |
|
|
|
| 994 |
|
|
## User Declared output files
|
| 995 |
slacapra |
1.54 |
for out in (self.output_file+self.output_file_sandbox):
|
| 996 |
slacapra |
1.1 |
n_out = nj + 1
|
| 997 |
|
|
out_box.append(self.numberFile_(out,str(n_out)))
|
| 998 |
|
|
return out_box
|
| 999 |
|
|
|
| 1000 |
|
|
def prepareSteeringCards(self):
|
| 1001 |
|
|
"""
|
| 1002 |
|
|
Make initial modifications of the user's steering card file.
|
| 1003 |
|
|
"""
|
| 1004 |
|
|
return
|
| 1005 |
|
|
|
| 1006 |
|
|
def wsRenameOutput(self, nj):
|
| 1007 |
|
|
"""
|
| 1008 |
|
|
Returns part of a job script which renames the produced files.
|
| 1009 |
|
|
"""
|
| 1010 |
|
|
|
| 1011 |
|
|
txt = '\n'
|
| 1012 |
gutsche |
1.7 |
txt += '# directory content\n'
|
| 1013 |
|
|
txt += 'ls \n'
|
| 1014 |
slacapra |
1.54 |
|
| 1015 |
|
|
for fileWithSuffix in (self.output_file+self.output_file_sandbox):
|
| 1016 |
slacapra |
1.1 |
output_file_num = self.numberFile_(fileWithSuffix, '$NJob')
|
| 1017 |
|
|
txt += '\n'
|
| 1018 |
gutsche |
1.7 |
txt += '# check output file\n'
|
| 1019 |
slacapra |
1.1 |
txt += 'ls '+fileWithSuffix+'\n'
|
| 1020 |
fanzago |
1.18 |
txt += 'ls_result=$?\n'
|
| 1021 |
|
|
txt += 'if [ $ls_result -ne 0 ] ; then\n'
|
| 1022 |
|
|
txt += ' echo "ERROR: Problem with output file"\n'
|
| 1023 |
gutsche |
1.7 |
if common.scheduler.boss_scheduler_name == 'condor_g':
|
| 1024 |
|
|
txt += ' if [ $middleware == OSG ]; then \n'
|
| 1025 |
|
|
txt += ' echo "prepare dummy output file"\n'
|
| 1026 |
|
|
txt += ' echo "Processing of job output failed" > $RUNTIME_AREA/'+output_file_num+'\n'
|
| 1027 |
|
|
txt += ' fi \n'
|
| 1028 |
slacapra |
1.1 |
txt += 'else\n'
|
| 1029 |
|
|
txt += ' cp '+fileWithSuffix+' $RUNTIME_AREA/'+output_file_num+'\n'
|
| 1030 |
|
|
txt += 'fi\n'
|
| 1031 |
|
|
|
| 1032 |
gutsche |
1.7 |
txt += 'cd $RUNTIME_AREA\n'
|
| 1033 |
fanzago |
1.18 |
txt += 'cd $RUNTIME_AREA\n'
|
| 1034 |
gutsche |
1.3 |
### OLI_DANIELE
|
| 1035 |
|
|
txt += 'if [ $middleware == OSG ]; then\n'
|
| 1036 |
|
|
txt += ' cd $RUNTIME_AREA\n'
|
| 1037 |
|
|
txt += ' echo "Remove working directory: $WORKING_DIR"\n'
|
| 1038 |
|
|
txt += ' /bin/rm -rf $WORKING_DIR\n'
|
| 1039 |
|
|
txt += ' if [ -d $WORKING_DIR ] ;then\n'
|
| 1040 |
gutsche |
1.7 |
txt += ' echo "SET_EXE 60999 ==> OSG $WORKING_DIR could not be deleted on WN `hostname` after cleanup of WN"\n'
|
| 1041 |
|
|
txt += ' echo "JOB_EXIT_STATUS = 60999"\n'
|
| 1042 |
|
|
txt += ' echo "JobExitCode=60999" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1043 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 1044 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 1045 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 1046 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1047 |
gutsche |
1.3 |
txt += ' fi\n'
|
| 1048 |
|
|
txt += 'fi\n'
|
| 1049 |
|
|
txt += '\n'
|
| 1050 |
slacapra |
1.54 |
|
| 1051 |
|
|
file_list = ''
|
| 1052 |
|
|
## Add to filelist only files to be possibly copied to SE
|
| 1053 |
|
|
for fileWithSuffix in self.output_file:
|
| 1054 |
|
|
output_file_num = self.numberFile_(fileWithSuffix, '$NJob')
|
| 1055 |
|
|
file_list=file_list+output_file_num+' '
|
| 1056 |
|
|
file_list=file_list[:-1]
|
| 1057 |
|
|
txt += 'file_list="'+file_list+'"\n'
|
| 1058 |
|
|
|
| 1059 |
slacapra |
1.1 |
return txt
|
| 1060 |
|
|
|
| 1061 |
|
|
def numberFile_(self, file, txt):
|
| 1062 |
|
|
"""
|
| 1063 |
|
|
append _'txt' before last extension of a file
|
| 1064 |
|
|
"""
|
| 1065 |
|
|
p = string.split(file,".")
|
| 1066 |
|
|
# take away last extension
|
| 1067 |
|
|
name = p[0]
|
| 1068 |
|
|
for x in p[1:-1]:
|
| 1069 |
|
|
name=name+"."+x
|
| 1070 |
|
|
# add "_txt"
|
| 1071 |
|
|
if len(p)>1:
|
| 1072 |
|
|
ext = p[len(p)-1]
|
| 1073 |
|
|
result = name + '_' + txt + "." + ext
|
| 1074 |
|
|
else:
|
| 1075 |
|
|
result = name + '_' + txt
|
| 1076 |
|
|
|
| 1077 |
|
|
return result
|
| 1078 |
|
|
|
| 1079 |
slacapra |
1.63 |
def getRequirements(self, nj=[]):
|
| 1080 |
slacapra |
1.1 |
"""
|
| 1081 |
|
|
return job requirements to add to jdl files
|
| 1082 |
|
|
"""
|
| 1083 |
|
|
req = ''
|
| 1084 |
slacapra |
1.47 |
if self.version:
|
| 1085 |
slacapra |
1.10 |
req='Member("VO-cms-' + \
|
| 1086 |
slacapra |
1.47 |
self.version + \
|
| 1087 |
slacapra |
1.10 |
'", other.GlueHostApplicationSoftwareRunTimeEnvironment)'
|
| 1088 |
gutsche |
1.35 |
|
| 1089 |
|
|
req = req + ' && (other.GlueHostNetworkAdapterOutboundIP)'
|
| 1090 |
|
|
|
| 1091 |
slacapra |
1.1 |
return req
|
| 1092 |
gutsche |
1.3 |
|
| 1093 |
|
|
def configFilename(self):
|
| 1094 |
|
|
""" return the config filename """
|
| 1095 |
|
|
return self.name()+'.cfg'
|
| 1096 |
|
|
|
| 1097 |
|
|
### OLI_DANIELE
|
| 1098 |
|
|
def wsSetupCMSOSGEnvironment_(self):
|
| 1099 |
|
|
"""
|
| 1100 |
|
|
Returns part of a job script which is prepares
|
| 1101 |
|
|
the execution environment and which is common for all CMS jobs.
|
| 1102 |
|
|
"""
|
| 1103 |
|
|
txt = '\n'
|
| 1104 |
|
|
txt += ' echo "### SETUP CMS OSG ENVIRONMENT ###"\n'
|
| 1105 |
|
|
txt += ' if [ -f $GRID3_APP_DIR/cmssoft/cmsset_default.sh ] ;then\n'
|
| 1106 |
|
|
txt += ' # Use $GRID3_APP_DIR/cmssoft/cmsset_default.sh to setup cms software\n'
|
| 1107 |
|
|
txt += ' source $GRID3_APP_DIR/cmssoft/cmsset_default.sh '+self.version+'\n'
|
| 1108 |
mkirn |
1.40 |
txt += ' elif [ -f $OSG_APP/cmssoft/cms/cmsset_default.sh ] ;then\n'
|
| 1109 |
|
|
txt += ' # Use $OSG_APP/cmssoft/cms/cmsset_default.sh to setup cms software\n'
|
| 1110 |
|
|
txt += ' source $OSG_APP/cmssoft/cms/cmsset_default.sh '+self.version+'\n'
|
| 1111 |
gutsche |
1.3 |
txt += ' else\n'
|
| 1112 |
mkirn |
1.40 |
txt += ' echo "SET_CMS_ENV 10020 ==> ERROR $GRID3_APP_DIR/cmssoft/cmsset_default.sh and $OSG_APP/cmssoft/cms/cmsset_default.sh file not found"\n'
|
| 1113 |
gutsche |
1.3 |
txt += ' echo "JOB_EXIT_STATUS = 10020"\n'
|
| 1114 |
|
|
txt += ' echo "JobExitCode=10020" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1115 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 1116 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 1117 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 1118 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1119 |
gutsche |
1.7 |
txt += ' exit 1\n'
|
| 1120 |
gutsche |
1.3 |
txt += '\n'
|
| 1121 |
|
|
txt += ' echo "Remove working directory: $WORKING_DIR"\n'
|
| 1122 |
|
|
txt += ' cd $RUNTIME_AREA\n'
|
| 1123 |
|
|
txt += ' /bin/rm -rf $WORKING_DIR\n'
|
| 1124 |
|
|
txt += ' if [ -d $WORKING_DIR ] ;then\n'
|
| 1125 |
mkirn |
1.40 |
txt += ' echo "SET_CMS_ENV 10017 ==> OSG $WORKING_DIR could not be deleted on WN `hostname` after $GRID3_APP_DIR/cmssoft/cmsset_default.sh and $OSG_APP/cmssoft/cms/cmsset_default.sh file not found"\n'
|
| 1126 |
gutsche |
1.7 |
txt += ' echo "JOB_EXIT_STATUS = 10017"\n'
|
| 1127 |
|
|
txt += ' echo "JobExitCode=10017" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1128 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 1129 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 1130 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 1131 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1132 |
gutsche |
1.3 |
txt += ' fi\n'
|
| 1133 |
|
|
txt += '\n'
|
| 1134 |
gutsche |
1.7 |
txt += ' exit 1\n'
|
| 1135 |
gutsche |
1.3 |
txt += ' fi\n'
|
| 1136 |
|
|
txt += '\n'
|
| 1137 |
|
|
txt += ' echo "SET_CMS_ENV 0 ==> setup cms environment ok"\n'
|
| 1138 |
|
|
txt += ' echo " END SETUP CMS OSG ENVIRONMENT "\n'
|
| 1139 |
|
|
|
| 1140 |
|
|
return txt
|
| 1141 |
|
|
|
| 1142 |
|
|
### OLI_DANIELE
|
| 1143 |
|
|
def wsSetupCMSLCGEnvironment_(self):
|
| 1144 |
|
|
"""
|
| 1145 |
|
|
Returns part of a job script which is prepares
|
| 1146 |
|
|
the execution environment and which is common for all CMS jobs.
|
| 1147 |
|
|
"""
|
| 1148 |
|
|
txt = ' \n'
|
| 1149 |
|
|
txt += ' echo " ### SETUP CMS LCG ENVIRONMENT ### "\n'
|
| 1150 |
|
|
txt += ' if [ ! $VO_CMS_SW_DIR ] ;then\n'
|
| 1151 |
|
|
txt += ' echo "SET_CMS_ENV 10031 ==> ERROR CMS software dir not found on WN `hostname`"\n'
|
| 1152 |
|
|
txt += ' echo "JOB_EXIT_STATUS = 10031" \n'
|
| 1153 |
|
|
txt += ' echo "JobExitCode=10031" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1154 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 1155 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 1156 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 1157 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1158 |
gutsche |
1.7 |
txt += ' exit 1\n'
|
| 1159 |
gutsche |
1.3 |
txt += ' else\n'
|
| 1160 |
|
|
txt += ' echo "Sourcing environment... "\n'
|
| 1161 |
|
|
txt += ' if [ ! -s $VO_CMS_SW_DIR/cmsset_default.sh ] ;then\n'
|
| 1162 |
|
|
txt += ' echo "SET_CMS_ENV 10020 ==> ERROR cmsset_default.sh file not found into dir $VO_CMS_SW_DIR"\n'
|
| 1163 |
|
|
txt += ' echo "JOB_EXIT_STATUS = 10020"\n'
|
| 1164 |
|
|
txt += ' echo "JobExitCode=10020" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1165 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 1166 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 1167 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 1168 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1169 |
gutsche |
1.7 |
txt += ' exit 1\n'
|
| 1170 |
gutsche |
1.3 |
txt += ' fi\n'
|
| 1171 |
|
|
txt += ' echo "sourcing $VO_CMS_SW_DIR/cmsset_default.sh"\n'
|
| 1172 |
|
|
txt += ' source $VO_CMS_SW_DIR/cmsset_default.sh\n'
|
| 1173 |
|
|
txt += ' result=$?\n'
|
| 1174 |
|
|
txt += ' if [ $result -ne 0 ]; then\n'
|
| 1175 |
|
|
txt += ' echo "SET_CMS_ENV 10032 ==> ERROR problem sourcing $VO_CMS_SW_DIR/cmsset_default.sh"\n'
|
| 1176 |
|
|
txt += ' echo "JOB_EXIT_STATUS = 10032"\n'
|
| 1177 |
|
|
txt += ' echo "JobExitCode=10032" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1178 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 1179 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 1180 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 1181 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1182 |
gutsche |
1.7 |
txt += ' exit 1\n'
|
| 1183 |
gutsche |
1.3 |
txt += ' fi\n'
|
| 1184 |
|
|
txt += ' fi\n'
|
| 1185 |
|
|
txt += ' \n'
|
| 1186 |
|
|
txt += ' string=`cat /etc/redhat-release`\n'
|
| 1187 |
|
|
txt += ' echo $string\n'
|
| 1188 |
|
|
txt += ' if [[ $string = *alhalla* ]]; then\n'
|
| 1189 |
|
|
txt += ' echo "SCRAM_ARCH= $SCRAM_ARCH"\n'
|
| 1190 |
|
|
txt += ' elif [[ $string = *Enterprise* ]] || [[ $string = *cientific* ]]; then\n'
|
| 1191 |
|
|
txt += ' export SCRAM_ARCH=slc3_ia32_gcc323\n'
|
| 1192 |
|
|
txt += ' echo "SCRAM_ARCH= $SCRAM_ARCH"\n'
|
| 1193 |
|
|
txt += ' else\n'
|
| 1194 |
gutsche |
1.7 |
txt += ' echo "SET_CMS_ENV 10033 ==> ERROR OS unknown, LCG environment not initialized"\n'
|
| 1195 |
gutsche |
1.3 |
txt += ' echo "JOB_EXIT_STATUS = 10033"\n'
|
| 1196 |
|
|
txt += ' echo "JobExitCode=10033" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1197 |
|
|
txt += ' dumpStatus $RUNTIME_AREA/$repo\n'
|
| 1198 |
gutsche |
1.13 |
txt += ' rm -f $RUNTIME_AREA/$repo \n'
|
| 1199 |
|
|
txt += ' echo "MonitorJobID=`echo $MonitorJobID`" | tee -a $RUNTIME_AREA/$repo \n'
|
| 1200 |
|
|
txt += ' echo "MonitorID=`echo $MonitorID`" | tee -a $RUNTIME_AREA/$repo\n'
|
| 1201 |
gutsche |
1.7 |
txt += ' exit 1\n'
|
| 1202 |
gutsche |
1.3 |
txt += ' fi\n'
|
| 1203 |
|
|
txt += ' echo "SET_CMS_ENV 0 ==> setup cms environment ok"\n'
|
| 1204 |
|
|
txt += ' echo "### END SETUP CMS LCG ENVIRONMENT ###"\n'
|
| 1205 |
|
|
return txt
|
| 1206 |
gutsche |
1.5 |
|
| 1207 |
|
|
def setParam_(self, param, value):
|
| 1208 |
|
|
self._params[param] = value
|
| 1209 |
|
|
|
| 1210 |
|
|
def getParams(self):
|
| 1211 |
|
|
return self._params
|
| 1212 |
gutsche |
1.8 |
|
| 1213 |
|
|
def setTaskid_(self):
|
| 1214 |
|
|
self._taskId = self.cfg_params['taskId']
|
| 1215 |
|
|
|
| 1216 |
|
|
def getTaskid(self):
|
| 1217 |
|
|
return self._taskId
|
| 1218 |
gutsche |
1.35 |
|
| 1219 |
|
|
#######################################################################
|
| 1220 |
|
|
def uniquelist(self, old):
|
| 1221 |
|
|
"""
|
| 1222 |
|
|
remove duplicates from a list
|
| 1223 |
|
|
"""
|
| 1224 |
|
|
nd={}
|
| 1225 |
|
|
for e in old:
|
| 1226 |
|
|
nd[e]=0
|
| 1227 |
|
|
return nd.keys()
|
