| 1 |
paus |
1.2 |
#!/usr/bin/env python
|
| 2 |
|
|
#---------------------------------------------------------------------------------------------------
|
| 3 |
|
|
# Script to test, create and submit one complete production task
|
| 4 |
|
|
#
|
| 5 |
|
|
# Complete refers here to the proper preparation of the ultimate storage location (a storge element
|
| 6 |
|
|
# with a given storage path etc.), the submission of a test job to evaluate timing and data output
|
| 7 |
|
|
# size and finally the creation of the job configurations and the submission of the jobs to the grid
|
| 8 |
|
|
# sites. The grid is accessed via crab tools.
|
| 9 |
|
|
#
|
| 10 |
|
|
# While the script is pretty complete it has become a bit long and messy. Some cleanup will
|
| 11 |
|
|
# hopefully soon be performed.
|
| 12 |
|
|
#
|
| 13 |
|
|
# Author: C.Paus (July 1, 2008)
|
| 14 |
|
|
#---------------------------------------------------------------------------------------------------
|
| 15 |
|
|
import os,sys,getopt,re,string
|
| 16 |
|
|
import task
|
| 17 |
|
|
|
| 18 |
|
|
#===================================================================================================
|
| 19 |
|
|
def searchReplace(line,mitCfg,version,mitDataset, \
|
| 20 |
|
|
cmsDataset='X',cmsswPy='X',dbs='X',sched='X',blacklist='X',skpEvts='X'):
|
| 21 |
|
|
# compile search and replacement sequences
|
| 22 |
|
|
pCmsDset = re.compile('XX-CMSDATASET-XX')
|
| 23 |
|
|
pMitDset = re.compile('XX-MITDATASET-XX')
|
| 24 |
|
|
pCmsswPy = re.compile('XX-CMSSWPY-XX')
|
| 25 |
|
|
pMitCfg = re.compile('XX-MITCFG-XX')
|
| 26 |
|
|
pMitVers = re.compile('XX-MITVERSION-XX')
|
| 27 |
|
|
pDbs = re.compile('XX-DBS-XX')
|
| 28 |
|
|
pSched = re.compile('XX-SCHED-XX')
|
| 29 |
|
|
pBlacklist = re.compile('XX-BLACKLIST-XX')
|
| 30 |
|
|
pSkpEvts = re.compile('XX-SKIPEVTS-XX')
|
| 31 |
|
|
# perform all search and replaces
|
| 32 |
|
|
line = pCmsDset .sub(cmsDataset, line);
|
| 33 |
|
|
line = pMitDset .sub(mitDataset, line);
|
| 34 |
|
|
line = pCmsswPy .sub(cmsswPy, line);
|
| 35 |
|
|
line = pMitCfg .sub(mitCfg, line);
|
| 36 |
|
|
line = pMitVers .sub(version, line);
|
| 37 |
|
|
line = pDbs .sub(dbs, line);
|
| 38 |
|
|
line = pSched .sub(sched, line);
|
| 39 |
|
|
line = pBlacklist.sub(blacklist, line);
|
| 40 |
|
|
return line
|
| 41 |
|
|
|
| 42 |
|
|
#===================================================================================================
|
| 43 |
|
|
def adjustCfg(line,nevents,crabId):
|
| 44 |
|
|
# compile search and replacement sequences
|
| 45 |
|
|
pNevents = re.compile('XX-NEVENTS-XX')
|
| 46 |
|
|
pCrabId = re.compile('XX-CRABID-XX')
|
| 47 |
|
|
# perform all search and replaces
|
| 48 |
|
|
line = pNevents .sub(str(nevents),line);
|
| 49 |
|
|
line = pCrabId .sub(crabId, line);
|
| 50 |
|
|
return line
|
| 51 |
|
|
|
| 52 |
|
|
#===================================================================================================
|
| 53 |
|
|
def findStoragePath(mitCfg,version,mitDataset):
|
| 54 |
|
|
# find the forseen storage place
|
| 55 |
|
|
cmd = 'grep ^storage_path ' + mitCfg + '/' + version + '/crab.cfg'
|
| 56 |
|
|
for file in os.popen(cmd).readlines():
|
| 57 |
|
|
line = file[:-1]
|
| 58 |
|
|
line = searchReplace(line,mitCfg,version,mitDataset);
|
| 59 |
|
|
# decode the storage directory name
|
| 60 |
|
|
names = line.split("=")
|
| 61 |
|
|
names = names[1:]
|
| 62 |
|
|
storagePath = "=".join(names)
|
| 63 |
|
|
storagePath = re.sub("\s", "",storagePath)
|
| 64 |
|
|
cmd = 'grep ^user_remote_dir ' + mitCfg + '/' + version + '/crab.cfg'
|
| 65 |
|
|
for file in os.popen(cmd).readlines():
|
| 66 |
|
|
line = file[:-1]
|
| 67 |
|
|
line = searchReplace(line,mitCfg,version,mitDataset);
|
| 68 |
|
|
# decode the storage directory name
|
| 69 |
|
|
names = line.split(" ")
|
| 70 |
|
|
storagePath += names[-1]
|
| 71 |
|
|
return storagePath
|
| 72 |
|
|
|
| 73 |
|
|
#===================================================================================================
|
| 74 |
|
|
def createDirCern(storagePath):
|
| 75 |
|
|
# check whether path exists
|
| 76 |
|
|
cmd = 'rfdir ' + storagePath + ' >& /dev/null'
|
| 77 |
|
|
status = os.system(cmd)
|
| 78 |
|
|
|
| 79 |
|
|
# create it if missing
|
| 80 |
|
|
if status == 0:
|
| 81 |
|
|
print ' Castor directory exists: ' + storagePath + '\n --> Moving on.'
|
| 82 |
|
|
else:
|
| 83 |
|
|
print ' Castor directory needs to be created.'
|
| 84 |
|
|
cmd = 'rfmkdir -p ' + storagePath
|
| 85 |
|
|
status = os.system(cmd)
|
| 86 |
|
|
if status == 0:
|
| 87 |
|
|
print ' --> Created: ' + storagePath
|
| 88 |
|
|
|
| 89 |
|
|
# always set the permissions
|
| 90 |
|
|
cmd = 'rfchmod 777 ' + storagePath
|
| 91 |
|
|
status = os.system(cmd)
|
| 92 |
|
|
if status == 0:
|
| 93 |
|
|
print ' --> Set permissions: 777\n'
|
| 94 |
|
|
else:
|
| 95 |
|
|
print ' --> Setting permissions failed. EXIT now.\n'
|
| 96 |
|
|
sys.exit(1)
|
| 97 |
|
|
|
| 98 |
|
|
#===================================================================================================
|
| 99 |
|
|
def createDirGeneral(storageEle,storagePath):
|
| 100 |
|
|
# create all relevant subdirectories
|
| 101 |
|
|
f = storagePath.split('/') # splitting every '/'
|
| 102 |
|
|
storagePath2 = "/".join(f[:-1])
|
| 103 |
|
|
storagePath1 = "/".join(f[:-2])
|
| 104 |
|
|
storagePath0 = "/".join(f[:-3])
|
| 105 |
|
|
|
| 106 |
|
|
# set the storage URL
|
| 107 |
|
|
storageUrl = 'srm://' + storageEle + ':8443' + storagePath
|
| 108 |
|
|
|
| 109 |
|
|
# check whether path exists
|
| 110 |
|
|
cmd = 'srmls ' + storageUrl + ' >& /dev/null'
|
| 111 |
|
|
status = os.system(cmd)
|
| 112 |
|
|
|
| 113 |
|
|
# create it only if missing
|
| 114 |
|
|
if status == 0:
|
| 115 |
|
|
print ' '
|
| 116 |
|
|
print ' Directory already found.... moving on.'
|
| 117 |
|
|
|
| 118 |
|
|
else:
|
| 119 |
|
|
# create all relevant directories
|
| 120 |
|
|
cmd = 'srmmkdir srm://' + storageEle + ':8443' + storagePath0 + ' >& /dev/null'
|
| 121 |
|
|
print ' srmmkdir: ' + cmd
|
| 122 |
|
|
status = os.system(cmd)
|
| 123 |
|
|
print ' srmmkdir: status %d'%(status)
|
| 124 |
|
|
cmd = 'srmmkdir srm://' + storageEle + ':8443' + storagePath1 + ' >& /dev/null'
|
| 125 |
|
|
print ' srmmkdir: ' + cmd
|
| 126 |
|
|
status = os.system(cmd)
|
| 127 |
|
|
print ' srmmkdir: status %d'%(status)
|
| 128 |
|
|
cmd = 'srmmkdir srm://' + storageEle + ':8443' + storagePath2 + ' >& /dev/null'
|
| 129 |
|
|
print ' srmmkdir: ' + cmd
|
| 130 |
|
|
status = os.system(cmd)
|
| 131 |
|
|
print ' srmmkdir: status %d'%(status)
|
| 132 |
|
|
|
| 133 |
|
|
# create the main storage directory
|
| 134 |
|
|
cmd = 'srmmkdir srm://' + storageEle + ':8443' + storagePath + ' >& /dev/null'
|
| 135 |
|
|
print ' srmmkdir: ' + cmd
|
| 136 |
|
|
status = os.system(cmd)
|
| 137 |
|
|
print ' srmmkdir: status %d'%(status)
|
| 138 |
|
|
if status == 0:
|
| 139 |
|
|
print ' '
|
| 140 |
|
|
print ' Directory was created.'
|
| 141 |
|
|
else:
|
| 142 |
|
|
print ' '
|
| 143 |
|
|
print ' '
|
| 144 |
|
|
print ' '
|
| 145 |
|
|
print ' ERROR - failed to create the Storage Area.'
|
| 146 |
|
|
print ' '
|
| 147 |
|
|
print ' '
|
| 148 |
|
|
|
| 149 |
|
|
print ' Check permissions with: srmls -l -count=1 ' + storageUrl + '\n'
|
| 150 |
|
|
cmd = 'srmls -l -count=1 ' + storageUrl + ' | grep GroupPermission | grep RWX'
|
| 151 |
|
|
status = -1
|
| 152 |
|
|
for line in os.popen(cmd).readlines(): # run command
|
| 153 |
|
|
line = line[:-1]
|
| 154 |
|
|
print " Permissions? " + line
|
| 155 |
|
|
status = 0
|
| 156 |
|
|
#status = os.system(cmd)
|
| 157 |
|
|
print ' '
|
| 158 |
|
|
|
| 159 |
|
|
if status != 0:
|
| 160 |
|
|
print ' --> Directory creation or permissions failed. EXIT now.\n'
|
| 161 |
|
|
sys.exit(1)
|
| 162 |
|
|
|
| 163 |
|
|
#===================================================================================================
|
| 164 |
|
|
# Main starts here
|
| 165 |
|
|
#===================================================================================================
|
| 166 |
|
|
# Define string to explain usage of the script
|
| 167 |
|
|
usage = "Usage: submit.py --cmsDataset=<name> | --mitDataset=<name>\n"
|
| 168 |
|
|
usage += " --cmssw=<name>\n"
|
| 169 |
|
|
usage += " --mitCfg=<name>\n"
|
| 170 |
|
|
usage += " --version=<version>\n"
|
| 171 |
|
|
usage += " --dbs=<name>\n"
|
| 172 |
|
|
usage += " --sched=<name>\n"
|
| 173 |
|
|
usage += " --blacklist=<name>\n"
|
| 174 |
|
|
usage += " --nSubmit=<submittedJobs>\n"
|
| 175 |
|
|
usage += " --skipEvents=<'nRunX:nEventY','nRunXX:nEventYY',...>\n"
|
| 176 |
|
|
usage += " --complete\n"
|
| 177 |
|
|
usage += " --testJob\n"
|
| 178 |
|
|
usage += " --noTestJob\n"
|
| 179 |
|
|
usage += " --help\n"
|
| 180 |
|
|
|
| 181 |
|
|
# Define the valid options which can be specified and check out the command line
|
| 182 |
|
|
valid = ['cmsDataset=','mitDataset=','cmssw=','mitCfg=','version=','dbs=','sched=','blacklist=',
|
| 183 |
|
|
'nSubmit=','skipEvents=','complete','testJob','noTestJob','test','help']
|
| 184 |
|
|
try:
|
| 185 |
|
|
opts, args = getopt.getopt(sys.argv[1:], "", valid)
|
| 186 |
|
|
except getopt.GetoptError, ex:
|
| 187 |
|
|
print usage
|
| 188 |
|
|
print str(ex)
|
| 189 |
|
|
sys.exit(1)
|
| 190 |
|
|
|
| 191 |
|
|
# --------------------------------------------------------------------------------------------------
|
| 192 |
|
|
# Get all parameters for the production
|
| 193 |
|
|
# --------------------------------------------------------------------------------------------------
|
| 194 |
|
|
# Set defaults for each option
|
| 195 |
|
|
cmsDataset = None
|
| 196 |
|
|
mitDataset = None
|
| 197 |
|
|
cmd = "date +crab_0_%y%m%d_%H%M%S"
|
| 198 |
|
|
for line in os.popen(cmd).readlines(): # run command
|
| 199 |
|
|
line = line[:-1]
|
| 200 |
|
|
crabId = line
|
| 201 |
|
|
print "\n This job will be CrabId: " + crabId + "\n"
|
| 202 |
|
|
|
| 203 |
|
|
cmssw = "cmssw"
|
| 204 |
|
|
mitCfg = "filler"
|
| 205 |
|
|
version = "013"
|
| 206 |
|
|
#dbs = "https://cmsdbsprod.cern.ch:8443/cms_dbs_prod_global/servlet/DBSServlet"
|
| 207 |
|
|
dbs = "http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet"
|
| 208 |
|
|
sched = "glite"
|
| 209 |
|
|
blacklist = ""
|
| 210 |
|
|
nSubmit = -1
|
| 211 |
|
|
skpEvts = ''
|
| 212 |
|
|
complete = 0
|
| 213 |
|
|
noTestJob = 0
|
| 214 |
|
|
testJob = 0
|
| 215 |
|
|
test = 0
|
| 216 |
|
|
|
| 217 |
|
|
# Read new values from the command line
|
| 218 |
|
|
for opt, arg in opts:
|
| 219 |
|
|
if opt == "--help":
|
| 220 |
|
|
print usage
|
| 221 |
|
|
sys.exit(0)
|
| 222 |
|
|
if opt == "--cmsDataset":
|
| 223 |
|
|
cmsDataset = arg
|
| 224 |
|
|
if opt == "--mitDataset":
|
| 225 |
|
|
mitDataset = arg
|
| 226 |
|
|
if opt == "--cmssw":
|
| 227 |
|
|
cmssw = arg
|
| 228 |
|
|
if opt == "--mitCfg":
|
| 229 |
|
|
mitCfg = arg
|
| 230 |
|
|
if opt == "--version":
|
| 231 |
|
|
version = arg
|
| 232 |
|
|
if opt == "--dbs":
|
| 233 |
|
|
dbs = arg
|
| 234 |
|
|
if opt == "--sched":
|
| 235 |
|
|
sched = arg
|
| 236 |
|
|
if opt == "--blacklist":
|
| 237 |
|
|
blacklist = arg
|
| 238 |
|
|
if opt == "--nSubmit":
|
| 239 |
|
|
nSubmit = arg
|
| 240 |
|
|
if opt == "--skipEvents":
|
| 241 |
|
|
skpEvts = arg
|
| 242 |
|
|
if opt == "--complete":
|
| 243 |
|
|
complete = 1
|
| 244 |
|
|
if opt == "--noTestJob":
|
| 245 |
|
|
noTestJob = 1
|
| 246 |
|
|
testJob = 0
|
| 247 |
|
|
if opt == "--testJob":
|
| 248 |
|
|
testJob = 1
|
| 249 |
|
|
noTestJob = 0
|
| 250 |
|
|
if opt == "--test":
|
| 251 |
|
|
test = 1
|
| 252 |
|
|
|
| 253 |
|
|
# Make sure we have the right 'database' and the right config file
|
| 254 |
|
|
database = 'Productions'
|
| 255 |
|
|
cmsswPy = cmssw + '.py'
|
| 256 |
|
|
if cmssw != 'cmssw':
|
| 257 |
|
|
database += '.' + cmssw
|
| 258 |
|
|
|
| 259 |
|
|
# Deal with obvious problems
|
| 260 |
|
|
if cmsDataset == None and mitDataset == None:
|
| 261 |
|
|
cmd = "--cmsDataset | --mitDataset options not provided. One of them is required."
|
| 262 |
|
|
raise RuntimeError, cmd
|
| 263 |
|
|
|
| 264 |
|
|
crabFile = mitCfg + '/' + version + '/' + 'crab.cfg'
|
| 265 |
|
|
if not os.path.exists(crabFile):
|
| 266 |
|
|
cmd = "Crab file not found: %s" % crabFile
|
| 267 |
|
|
raise RuntimeError, cmd
|
| 268 |
|
|
cmsswFile = mitCfg + '/' + version + '/' + cmsswPy
|
| 269 |
|
|
if not os.path.exists(cmsswFile):
|
| 270 |
|
|
cmd = "Cmssw file not found: %s" % cmsswFile
|
| 271 |
|
|
cmd = " XXXX ERROR no valid configuration found XXXX"
|
| 272 |
|
|
raise RuntimeError, cmd
|
| 273 |
|
|
|
| 274 |
|
|
# Resolve the other mitCfg parameters from the configuration file
|
| 275 |
|
|
cmd = 'cat ' + mitCfg + '/' + version + '/' + database
|
| 276 |
|
|
|
| 277 |
|
|
join = 0
|
| 278 |
|
|
fullLine = ""
|
| 279 |
|
|
bSlash = "\\";
|
| 280 |
|
|
for line in os.popen(cmd).readlines(): # run command
|
| 281 |
|
|
line = line[:-1]
|
| 282 |
|
|
# get ride of empty or commented lines
|
| 283 |
|
|
if line == '' or line[0] == '#':
|
| 284 |
|
|
continue
|
| 285 |
|
|
|
| 286 |
|
|
# join lines
|
| 287 |
|
|
if join == 1:
|
| 288 |
|
|
fullLine += line
|
| 289 |
|
|
else:
|
| 290 |
|
|
fullLine = line
|
| 291 |
|
|
|
| 292 |
|
|
# determine if finished or more is coming
|
| 293 |
|
|
if fullLine[-1] == bSlash:
|
| 294 |
|
|
join = 1
|
| 295 |
|
|
fullLine = fullLine[:-1]
|
| 296 |
|
|
else:
|
| 297 |
|
|
join = 0
|
| 298 |
|
|
# test whether there is a directory
|
| 299 |
|
|
names = fullLine.split() # splitting every blank
|
| 300 |
|
|
if names[0] == cmsDataset or names[1] == mitDataset:
|
| 301 |
|
|
cmsDataset = names[0] # CMS name of the dataset
|
| 302 |
|
|
mitDataset = names[1] # equivalent MIT name of the dataset
|
| 303 |
|
|
nevents = int(names[2]) # number of events to be used in the production
|
| 304 |
|
|
if names[4] != "-":
|
| 305 |
|
|
localPath = names[4]
|
| 306 |
|
|
print "\n Sample info from database %s for CMSSW config %s\n %s"\
|
| 307 |
|
|
%(database,cmsswPy,fullLine)
|
| 308 |
|
|
if len(names) == 6:
|
| 309 |
|
|
dbs = names[5]
|
| 310 |
|
|
dbs = 'https://cmsdbsprod.cern.ch:8443/cms_dbs_' + dbs + '/servlet/DBSServlet'
|
| 311 |
|
|
print ' dbs: ' + dbs + '\n'
|
| 312 |
|
|
else:
|
| 313 |
|
|
print ''
|
| 314 |
|
|
|
| 315 |
|
|
if mitDataset == None or cmsDataset == None:
|
| 316 |
|
|
print "ERROR - dataset not defined."
|
| 317 |
|
|
sys.exit(1)
|
| 318 |
|
|
|
| 319 |
|
|
# Prepare file based processing input
|
| 320 |
|
|
runFile = mitCfg + '/' + version + '/' + 'run.sh'
|
| 321 |
|
|
if not os.path.exists(runFile):
|
| 322 |
|
|
cmd = "Run file not found: %s" % runFile
|
| 323 |
|
|
raise RuntimeError, cmd
|
| 324 |
|
|
cmd = 'cp ' + runFile + ' ./'
|
| 325 |
|
|
os.system(cmd)
|
| 326 |
|
|
|
| 327 |
|
|
lfnFile = mitCfg + '/' + version + '/' + mitDataset + '.lfns'
|
| 328 |
|
|
if os.path.exists(lfnFile):
|
| 329 |
|
|
cmd = "Lfn file found: %s. This means someone already worked on this dataset." % lfnFile
|
| 330 |
|
|
#### raise RuntimeError, cmd
|
| 331 |
|
|
else:
|
| 332 |
|
|
cmd = 'input.py --option=lfn --dataset=' + cmsDataset + ' > ' + lfnFile
|
| 333 |
|
|
print ' Input: ' + cmd
|
| 334 |
|
|
os.system(cmd)
|
| 335 |
|
|
|
| 336 |
|
|
# Create the corresponding crab task
|
| 337 |
|
|
crabTask = task.Task(crabId,cmsDataset,mitCfg,version)
|
| 338 |
|
|
crabTask.storagePath = findStoragePath(mitCfg,version,mitDataset)
|
| 339 |
|
|
crabTask.loadAllLfns(lfnFile)
|
| 340 |
|
|
crabTask.loadCompletedLfns()
|
| 341 |
|
|
crabTask.createMissingLfns(lfnFile,lfnFile + '_' + crabTask.tag)
|
| 342 |
|
|
crabTask.createSubTasks(lfnFile + '_' + crabTask.tag)
|
| 343 |
|
|
cmd = 'cp ' + lfnFile + '_' + crabTask.tag + '_*' + ' ./'
|
| 344 |
|
|
os.system(cmd)
|
| 345 |
|
|
|
| 346 |
|
|
nevents = len(crabTask.lfns)
|
| 347 |
|
|
|
| 348 |
|
|
# Say what we do now
|
| 349 |
|
|
print ' Preparing dataset: ' + cmsDataset + ' [MIT: ' + mitDataset + ' with ' + str(nevents) + \
|
| 350 |
|
|
' total jobs(=lfns), nEvtsTotal: %d]' % crabTask.nTotalEvts
|
| 351 |
|
|
|
| 352 |
|
|
# --------------------------------------------------------------------------------------------------
|
| 353 |
|
|
# Prepare the config files
|
| 354 |
|
|
# --------------------------------------------------------------------------------------------------
|
| 355 |
|
|
# Cleaning up
|
| 356 |
|
|
cmd = "rm -f crab.cfg crab.cfg-Template cmssw.cfg cmssw.py"
|
| 357 |
|
|
os.system(cmd)
|
| 358 |
|
|
|
| 359 |
|
|
# Parse template input and write the crab configuration file
|
| 360 |
|
|
fileInput = open(crabFile,'r')
|
| 361 |
|
|
fileOutput = open("crab.cfg-Template",'w')
|
| 362 |
|
|
line = fileInput.readline()
|
| 363 |
|
|
while (line != ''):
|
| 364 |
|
|
if line[0] != '#':
|
| 365 |
|
|
line = searchReplace(line,mitCfg,version,mitDataset, \
|
| 366 |
|
|
cmsDataset,cmsswPy,dbs,sched,blacklist,skpEvts)
|
| 367 |
|
|
fileOutput.write(line)
|
| 368 |
|
|
line = fileInput.readline()
|
| 369 |
|
|
fileInput .close()
|
| 370 |
|
|
fileOutput.close()
|
| 371 |
|
|
|
| 372 |
|
|
# Parse template input and write the crab configuration file
|
| 373 |
|
|
fileInput = open(cmsswFile,'r')
|
| 374 |
|
|
fileOutput = open(cmsswPy,'w')
|
| 375 |
|
|
line = fileInput.readline()
|
| 376 |
|
|
while (line != ''):
|
| 377 |
|
|
if line[0] != '#':
|
| 378 |
|
|
line = searchReplace(line,mitCfg,version,mitDataset, \
|
| 379 |
|
|
cmsDataset,cmsswPy,dbs,sched,blacklist,skpEvts)
|
| 380 |
|
|
fileOutput.write(line)
|
| 381 |
|
|
line = fileInput.readline()
|
| 382 |
|
|
fileInput .close()
|
| 383 |
|
|
fileOutput.close()
|
| 384 |
|
|
|
| 385 |
|
|
# --------------------------------------------------------------------------------------------------
|
| 386 |
|
|
# Job creation and submission
|
| 387 |
|
|
# --------------------------------------------------------------------------------------------------
|
| 388 |
|
|
print '\n Using CMSSW version: ' + os.environ['CMSSW_VERSION']
|
| 389 |
|
|
print ' Using CRAB version: ' + os.environ['CRAB_VERS'] + '\n'
|
| 390 |
|
|
|
| 391 |
|
|
pattern1 = 'working directory'
|
| 392 |
|
|
pattern2 = 'job\(s\) can run on'
|
| 393 |
|
|
|
| 394 |
|
|
# Go through the crabTask and submit each subtask to the grid
|
| 395 |
|
|
for subTask in crabTask.subTasks:
|
| 396 |
|
|
|
| 397 |
|
|
tag = crabTask.tag + '_' + subTask.tag()
|
| 398 |
|
|
print ' Working on subtask: ' + tag
|
| 399 |
|
|
|
| 400 |
|
|
nJobsTotal = 0
|
| 401 |
|
|
crabIdCheck = ''
|
| 402 |
|
|
|
| 403 |
|
|
# adjust crab config
|
| 404 |
|
|
fileInput = open("crab.cfg-Template",'r')
|
| 405 |
|
|
fileOutput = open("crab.cfg",'w')
|
| 406 |
|
|
line = fileInput.readline()
|
| 407 |
|
|
while (line != ''):
|
| 408 |
|
|
if line[0] != '#':
|
| 409 |
|
|
line = adjustCfg(line,subTask.nSubTaskLfn,tag)
|
| 410 |
|
|
fileOutput.write(line)
|
| 411 |
|
|
line = fileInput.readline()
|
| 412 |
|
|
fileInput .close()
|
| 413 |
|
|
fileOutput.close()
|
| 414 |
|
|
|
| 415 |
|
|
# ----------------------------------------------------------------------------------------------
|
| 416 |
|
|
# Deal with storage element area
|
| 417 |
|
|
# ----------------------------------------------------------------------------------------------
|
| 418 |
|
|
# find the forseen storage place
|
| 419 |
|
|
cmd = 'grep ^storage_element crab.cfg-Template'
|
| 420 |
|
|
for file in os.popen(cmd).readlines(): # run command
|
| 421 |
|
|
line = file[:-1] # strip '\n'
|
| 422 |
|
|
# decode the storage element name
|
| 423 |
|
|
names = line.split("=") # splitting every '='
|
| 424 |
|
|
storageEle = names.pop()
|
| 425 |
|
|
storageEle = re.sub("\s", "",storageEle)
|
| 426 |
|
|
cmd = 'grep ^storage_path crab.cfg-Template'
|
| 427 |
|
|
for file in os.popen(cmd).readlines(): # run command
|
| 428 |
|
|
line = file[:-1] # strip '\n'
|
| 429 |
|
|
# decode the storage directory name
|
| 430 |
|
|
names = line.split("=") # splitting every '='
|
| 431 |
|
|
names = names[1:]
|
| 432 |
|
|
storagePath = "=".join(names)
|
| 433 |
|
|
storagePath = re.sub("\s", "",storagePath)
|
| 434 |
|
|
cmd = 'grep ^user_remote_dir crab.cfg-Template'
|
| 435 |
|
|
for file in os.popen(cmd).readlines(): # run command
|
| 436 |
|
|
line = file[:-1] # strip '\n'
|
| 437 |
|
|
# decode the storage directory name
|
| 438 |
|
|
names = line.split(" ") # splitting every '='
|
| 439 |
|
|
storagePath += names[-1]
|
| 440 |
|
|
storageUrl = 'srm://' + storageEle + ':8443' + storagePath
|
| 441 |
|
|
|
| 442 |
|
|
storagePath = adjustCfg(storagePath,subTask.nSubTaskLfn,tag)
|
| 443 |
|
|
storageUrl = adjustCfg(storageUrl, subTask.nSubTaskLfn,tag)
|
| 444 |
|
|
|
| 445 |
|
|
print ' StorageUrl: ' + storageUrl
|
| 446 |
|
|
|
| 447 |
|
|
# Create storage area
|
| 448 |
|
|
if test == 0: # for testing we do not create the directories
|
| 449 |
|
|
if (storageEle == 'srm.cern.ch'):
|
| 450 |
|
|
createDirCern(storagePath)
|
| 451 |
|
|
else:
|
| 452 |
|
|
createDirGeneral(storageEle,storagePath)
|
| 453 |
|
|
|
| 454 |
|
|
# cmd = "crab -create -debug 3 -USER.ui_working_dir=" + tag + " | tee forDaniele "
|
| 455 |
|
|
cmd = "crab -create -USER.ui_working_dir=" + tag
|
| 456 |
|
|
print ' -> ' + cmd
|
| 457 |
|
|
if test != 0:
|
| 458 |
|
|
cmd = 'echo ' + cmd
|
| 459 |
|
|
print ' ' + cmd
|
| 460 |
|
|
for line in os.popen(cmd).readlines(): # run command
|
| 461 |
|
|
line = line[:-1] # strip '\n'
|
| 462 |
|
|
print ' CRAB: ' + line
|
| 463 |
|
|
if re.search(pattern1,line):
|
| 464 |
|
|
f = line.split("/")
|
| 465 |
|
|
crabIdCheck = f[-2]
|
| 466 |
|
|
if re.search(pattern2,line):
|
| 467 |
|
|
f = line.split(" ")
|
| 468 |
|
|
nJobsTotal = f[1]
|
| 469 |
|
|
if nJobsTotal == '':
|
| 470 |
|
|
nJobsTotal = f[2]
|
| 471 |
|
|
# and report
|
| 472 |
|
|
print ' --> %s jobs created (%s).\n'%(nJobsTotal,tag)
|
| 473 |
|
|
|
| 474 |
|
|
# adjust arguments
|
| 475 |
|
|
cmd = 'input.py --db=' + lfnFile + '_' + tag + ' --option=xml --dataset=' + cmsDataset + \
|
| 476 |
|
|
' > ' + tag + '/share/arguments.xml'
|
| 477 |
|
|
print ' update arguments: ' + cmd
|
| 478 |
|
|
if test == 0:
|
| 479 |
|
|
os.system(cmd)
|
| 480 |
|
|
|
| 481 |
|
|
# loop through the file and determine the submission parameters
|
| 482 |
|
|
block = ''
|
| 483 |
|
|
blocks = []
|
| 484 |
|
|
idx = 0
|
| 485 |
|
|
minIdxs = []
|
| 486 |
|
|
maxIdxs = []
|
| 487 |
|
|
|
| 488 |
|
|
fileInput = open(lfnFile + '_' + tag,'r')
|
| 489 |
|
|
line = fileInput.readline()
|
| 490 |
|
|
while (line != ''):
|
| 491 |
|
|
line = line[:-1]
|
| 492 |
|
|
if line[0] != '#':
|
| 493 |
|
|
idx += 1
|
| 494 |
|
|
f = line.split()
|
| 495 |
|
|
block = f[0]
|
| 496 |
|
|
lfn = f[1]
|
| 497 |
|
|
nEvents = f[2]
|
| 498 |
|
|
if len(blocks) == 0 or block != blocks[-1]:
|
| 499 |
|
|
# new block found
|
| 500 |
|
|
blocks .append(block)
|
| 501 |
|
|
minIdxs.append(idx)
|
| 502 |
|
|
maxIdxs.append(idx)
|
| 503 |
|
|
else:
|
| 504 |
|
|
maxIdxs[-1] = idx
|
| 505 |
|
|
# keep the reading going
|
| 506 |
|
|
line = fileInput.readline()
|
| 507 |
|
|
fileInput .close()
|
| 508 |
|
|
|
| 509 |
|
|
# merge blocks together if they use the same sites
|
| 510 |
|
|
lastSites = ''
|
| 511 |
|
|
idx = 0
|
| 512 |
|
|
mergedIdx = 0
|
| 513 |
|
|
mergedBlocks = []
|
| 514 |
|
|
mergedSites = []
|
| 515 |
|
|
mergedMinIdxs = []
|
| 516 |
|
|
mergedMaxIdxs = []
|
| 517 |
|
|
print '\n Show the unmerged blocks:'
|
| 518 |
|
|
for block in blocks:
|
| 519 |
|
|
cmd = "sites.py --block=" + block
|
| 520 |
|
|
for line in os.popen(cmd).readlines(): # run command
|
| 521 |
|
|
line = line[:-1]
|
| 522 |
|
|
sites = line
|
| 523 |
|
|
print ' Block ' + block + ' process: %d to %d'%(minIdxs[idx],maxIdxs[idx]) + \
|
| 524 |
|
|
' at\n > ' + sites
|
| 525 |
|
|
# block with different sites found
|
| 526 |
|
|
if sites != lastSites:
|
| 527 |
|
|
lastSites = sites
|
| 528 |
|
|
mergedSites .append(sites)
|
| 529 |
|
|
mergedBlocks .append(blocks[idx])
|
| 530 |
|
|
mergedMinIdxs.append(minIdxs[idx])
|
| 531 |
|
|
mergedMaxIdxs.append(maxIdxs[idx])
|
| 532 |
|
|
mergedIdx =+ 1
|
| 533 |
|
|
else:
|
| 534 |
|
|
mergedMaxIdxs[mergedIdx-1] = maxIdxs[idx]
|
| 535 |
|
|
# last action in the loop: increment the unmerged blocks
|
| 536 |
|
|
idx += 1
|
| 537 |
|
|
|
| 538 |
|
|
# Show already what we will do
|
| 539 |
|
|
idx = 0
|
| 540 |
|
|
print '\n Show the merged blocks:'
|
| 541 |
|
|
for block in mergedBlocks:
|
| 542 |
|
|
print ' Merged Block ' + block + ' process: %d to %d'\
|
| 543 |
|
|
%(mergedMinIdxs[idx],mergedMaxIdxs[idx]) + \
|
| 544 |
|
|
' at\n > ' + mergedSites[idx]
|
| 545 |
|
|
# last action in the loop: increment the merged blocks
|
| 546 |
|
|
idx += 1
|
| 547 |
|
|
|
| 548 |
|
|
# perfrom the submission block by block (using the merged blocks of course)
|
| 549 |
|
|
nSubmission = len(mergedBlocks)
|
| 550 |
|
|
idx = 0
|
| 551 |
|
|
print '\n Submit the merged blocks:'
|
| 552 |
|
|
for block in mergedBlocks:
|
| 553 |
|
|
print ' Merged Block ' + block + ' process: %d to %d'\
|
| 554 |
|
|
%(mergedMinIdxs[idx],mergedMaxIdxs[idx]) + \
|
| 555 |
|
|
' at\n > ' + mergedSites[idx]
|
| 556 |
|
|
|
| 557 |
|
|
nSubmit = '%d-%d'%(mergedMinIdxs[idx],mergedMaxIdxs[idx])
|
| 558 |
|
|
if mergedMinIdxs[idx] == mergedMaxIdxs[idx]:
|
| 559 |
|
|
nSubmit = '%d,%d'%(mergedMinIdxs[idx],100000000)
|
| 560 |
|
|
cmd = 'crab -submit %s -continue %s -GRID.se_white_list=%s'%(nSubmit,tag,mergedSites[idx])
|
| 561 |
|
|
print ' ' + cmd
|
| 562 |
|
|
os.system(cmd)
|
| 563 |
|
|
# last action in the loop: increment the merged blocks
|
| 564 |
|
|
idx += 1
|
| 565 |
|
|
|
| 566 |
|
|
print ' Number of blocks submitted: %d' % nSubmission
|
| 567 |
|
|
|
| 568 |
|
|
print ' Done... keep watching it...'
|
| 569 |
|
|
sys.exit(0)
|
| 570 |
|
|
|
| 571 |
|
|
|
| 572 |
|
|
## if nSubmit != -1:
|
| 573 |
|
|
## cmd = 'crab -continue %s -submit %s'%(crabId,nSubmit)
|
| 574 |
|
|
## print 'SUBMIT.PY: ' + cmd
|
| 575 |
|
|
## status = os.system(cmd)
|
| 576 |
|
|
## if status == 0:
|
| 577 |
|
|
## print ' --> job submitted\n'
|
| 578 |
|
|
## else:
|
| 579 |
|
|
## leftOver = int(nJobsTotal)
|
| 580 |
|
|
## nSubmitted = 0
|
| 581 |
|
|
## nSubBatch = 80
|
| 582 |
|
|
## while (nSubmitted < int(nJobsTotal)):
|
| 583 |
|
|
## if leftOver < nSubBatch:
|
| 584 |
|
|
## nSubBatch = leftOver
|
| 585 |
|
|
##
|
| 586 |
|
|
## cmd = 'crab -continue %s -submit %d'%(crabId,nSubBatch)
|
| 587 |
|
|
## print 'SUBMIT.PY: ' + cmd
|
| 588 |
|
|
## status = os.system(cmd)
|
| 589 |
|
|
## if status == 0:
|
| 590 |
|
|
## print ' --> %d job submitted\n'%(nSubBatch)
|
| 591 |
|
|
## leftOver = leftOver - nSubBatch
|
| 592 |
|
|
## nSubmitted = nSubmitted + nSubBatch
|
| 593 |
|
|
##
|
| 594 |
|
|
## #print 'SUBMIT.PY: ' + cmd
|
| 595 |
|
|
## #status = os.system(cmd)
|
| 596 |
|
|
## #if status == 0:
|
| 597 |
|
|
## # print ' --> job submitted\n'
|
| 598 |
|
|
##
|
| 599 |
|
|
|
| 600 |
|
|
# --------------------------------------------------------------------------------------------------
|
| 601 |
|
|
# Run a test job to test the configuration and measure the expected output size
|
| 602 |
|
|
# --------------------------------------------------------------------------------------------------
|
| 603 |
|
|
if noTestJob == 0:
|
| 604 |
|
|
#-----------------------------------------------------------------------------------------------
|
| 605 |
|
|
# use a specific file for test
|
| 606 |
|
|
#-----------------------------------------------------------------------------------------------
|
| 607 |
|
|
# first check whether we already have a local file
|
| 608 |
|
|
print '\n Try to find an existing local file using "find"'
|
| 609 |
|
|
f = cmsDataset.split("/")
|
| 610 |
|
|
name = f[1]
|
| 611 |
|
|
vers = f[2]
|
| 612 |
|
|
tier = f[3]
|
| 613 |
|
|
file = ""
|
| 614 |
|
|
lfn = ""
|
| 615 |
|
|
cmd = 'find ./store/ -name ' + name
|
| 616 |
|
|
print 'Searching: ' + cmd
|
| 617 |
|
|
for line in os.popen(cmd).readlines(): # run command
|
| 618 |
|
|
file = line[:-1] # strip '\n'
|
| 619 |
|
|
|
| 620 |
|
|
# looks like there could be a file we found a directory, confirm
|
| 621 |
|
|
if file != "":
|
| 622 |
|
|
print '\n We have a directory ' + file + ' .. confirming'
|
| 623 |
|
|
file = file + '/' + tier
|
| 624 |
|
|
cmd = 'find ' + file + ' -name \*.root'
|
| 625 |
|
|
for line in os.popen(cmd).readlines(): # run command
|
| 626 |
|
|
if line.find(vers):
|
| 627 |
|
|
file = line[:-1] # strip '\n'
|
| 628 |
|
|
lfn = file[1:]
|
| 629 |
|
|
|
| 630 |
|
|
if os.path.exists(file):
|
| 631 |
|
|
print ' moving on with locally found file: \n' + ' ' + file
|
| 632 |
|
|
else:
|
| 633 |
|
|
print ' no local file found'
|
| 634 |
|
|
file = ""
|
| 635 |
|
|
lfn = ""
|
| 636 |
|
|
|
| 637 |
|
|
# now try to see whether we can find a file to download
|
| 638 |
|
|
if file == "":
|
| 639 |
|
|
cmd = './bin/findLfn.py --input=' + cmsDataset + ' | grep /store/'
|
| 640 |
|
|
print '\n Find an LFN to download: ' + cmd
|
| 641 |
|
|
for line in os.popen(cmd).readlines(): # run command
|
| 642 |
|
|
if line.find("/store") != -1:
|
| 643 |
|
|
lfn = line[:-1] # strip '\n'
|
| 644 |
|
|
break
|
| 645 |
|
|
if lfn == "":
|
| 646 |
|
|
print "\n WARNING: No file found, continue assuming it is a simulation job.\n\n"
|
| 647 |
|
|
else:
|
| 648 |
|
|
print ' --> LFN: ' + lfn
|
| 649 |
|
|
file = '.' + lfn
|
| 650 |
|
|
|
| 651 |
|
|
if os.path.exists(file):
|
| 652 |
|
|
print ' --> File already exists: ' + file
|
| 653 |
|
|
else:
|
| 654 |
|
|
cmd = './bin/downloadLfn.py ' + lfn
|
| 655 |
|
|
print ' --> downloading: ' + cmd
|
| 656 |
|
|
status = os.system(cmd)
|
| 657 |
|
|
if status != 0:
|
| 658 |
|
|
print ' ERROR - failed to copy LFN. EXIT now!'
|
| 659 |
|
|
sys.exit(1)
|
| 660 |
|
|
|
| 661 |
|
|
# Parse template input and adjust the input file to the newly copied lfn
|
| 662 |
|
|
fileInput = open(cmsswPy,'r')
|
| 663 |
|
|
fileOutput = open("test-"+cmsswPy,'w')
|
| 664 |
|
|
line = fileInput.readline()
|
| 665 |
|
|
while (line != ''):
|
| 666 |
|
|
if line.find("file:") != -1:
|
| 667 |
|
|
line = '"file:' + lfn[1:] + '"\n'
|
| 668 |
|
|
fileOutput.write(line)
|
| 669 |
|
|
line = fileInput.readline()
|
| 670 |
|
|
fileInput .close()
|
| 671 |
|
|
fileOutput.close()
|
| 672 |
|
|
|
| 673 |
|
|
# Setting the number of events (hard coded in the file so far)
|
| 674 |
|
|
nTryEvts = 1000.
|
| 675 |
|
|
|
| 676 |
|
|
print '\n --> Please wait, running a test job now! Should be short (trying %.0f'%nTryEvts + \
|
| 677 |
|
|
' evts). Check log: cmssw.log'
|
| 678 |
|
|
cmd = 'rm -f cmssw.log; /usr/bin/time --format "%e %U %S" cmsRun test-' + cmsswPy + \
|
| 679 |
|
|
' >& cmssw.log'
|
| 680 |
|
|
print ' CMD: ' + cmd
|
| 681 |
|
|
status = os.system(cmd)
|
| 682 |
|
|
|
| 683 |
|
|
cmd = 'tail -1 cmssw.log'
|
| 684 |
|
|
for file in os.popen(cmd).readlines(): # run command
|
| 685 |
|
|
line = file[:-1] # strip '\n'
|
| 686 |
|
|
f = line.split() # splitting every blank
|
| 687 |
|
|
rtime = float(f[0]) # wall clock time
|
| 688 |
|
|
utime = float(f[1]) # user time
|
| 689 |
|
|
stime = float(f[2]) # system time
|
| 690 |
|
|
|
| 691 |
|
|
nEvtsTest = 1000
|
| 692 |
|
|
cmd = 'grep \'Begin processing\' cmssw.log | tail -1'
|
| 693 |
|
|
for file in os.popen(cmd).readlines(): # run command
|
| 694 |
|
|
line = file[:-1] # strip '\n'
|
| 695 |
|
|
# test whether there is a directory
|
| 696 |
|
|
f = line.split() # splitting every blank
|
| 697 |
|
|
nEvtsTest = f[3] # this is the number of records processed
|
| 698 |
|
|
nEvtsTest = int(nEvtsTest[:-2]) # strip 'th'
|
| 699 |
|
|
cmd = 'ls -s ' + mitDataset + '*.root'
|
| 700 |
|
|
size = 0
|
| 701 |
|
|
for file in os.popen(cmd).readlines(): # run command
|
| 702 |
|
|
line = file[:-1] # strip '\n'
|
| 703 |
|
|
f = line.split() # splitting every blank
|
| 704 |
|
|
size += int(f[0])/1000. # size in MB
|
| 705 |
|
|
|
| 706 |
|
|
cmd = 'tail -1 cmssw.log'
|
| 707 |
|
|
for file in os.popen(cmd).readlines(): # run command
|
| 708 |
|
|
line = file[:-1] # strip '\n'
|
| 709 |
|
|
# get total, user and system times
|
| 710 |
|
|
names = line.split() # splitting every blank
|
| 711 |
|
|
|
| 712 |
|
|
if nEvtsTest != nTryEvts:
|
| 713 |
|
|
print ' WARNING - Instead of %f did %d'%(nTryEvts,nEvtsTest)
|
| 714 |
|
|
|
| 715 |
|
|
print ' '
|
| 716 |
|
|
print ' Number of test events produced: %d'%nEvtsTest
|
| 717 |
|
|
print ' File size for all events: %.2f MB'%size
|
| 718 |
|
|
print ' Processing time for all events: %.2f secs (u: %.2f s: %.2f)'%(rtime,utime,stime)
|
| 719 |
|
|
print ' '
|
| 720 |
|
|
print ' --> 1 event == %.2f secs'%(rtime/nEvtsTest)
|
| 721 |
|
|
print ' --> 1.00 GB == %d events'%(nEvtsTest/size*1024.)
|
| 722 |
|
|
print ' --> %.2f GB == %d events'%(nevents/(nEvtsTest/size*1024.),nevents)
|
| 723 |
|
|
print ' '
|
| 724 |
|
|
|
| 725 |
|
|
if testJob == 1:
|
| 726 |
|
|
print '\n Test job finished, stopping now.\n'
|
| 727 |
|
|
sys.exit(0)
|
| 728 |
|
|
|
| 729 |
|
|
|
| 730 |
|
|
## # are we just completing an existing production? and is there something to complete?
|
| 731 |
|
|
## if complete == 1:
|
| 732 |
|
|
## f = storagePath.split('=')
|
| 733 |
|
|
## rfDir = f[-1]
|
| 734 |
|
|
## #cmd = 'castorInventory.py --nJobs=%s %s | grep Missing'%(nJobsTotal,rfDir)
|
| 735 |
|
|
## cmd = 'castorInventory.py --nJobs=%s %s | grep Missing'%(nJobsTotal,storageUrl)
|
| 736 |
|
|
## #print ' CMD: ' + cmd
|
| 737 |
|
|
## for line in os.popen(cmd).readlines(): # run command
|
| 738 |
|
|
## line = line[:-1] # strip '\n'
|
| 739 |
|
|
## f = line.split(':')
|
| 740 |
|
|
## nSubmit = f[1].strip()
|
| 741 |
|
|
## f = nSubmit.split(',')
|
| 742 |
|
|
## if len(f) == 0 or nSubmit == '':
|
| 743 |
|
|
## print ' No more jobs left it seems, nSubmit=' + nSubmit
|
| 744 |
|
|
## cmd = 'rm -rf ' + crabId
|
| 745 |
|
|
## print ' Cleanup: ' + cmd + '\n\n'
|
| 746 |
|
|
## status = os.system(cmd)
|
| 747 |
|
|
## sys.exit(0)
|
| 748 |
|
|
## elif len(f) == 1:
|
| 749 |
|
|
## nInvalid = str(int(nJobsTotal) + 1000)
|
| 750 |
|
|
## print ' One more jobs left, careful, adjusted, nSubmit=' + nSubmit
|
| 751 |
|
|
## nSubmit = nSubmit + ',' + nInvalid
|
| 752 |
|
|
##
|
| 753 |
|
|
## ### nSubmit = ",".join(f[:-1])
|
| 754 |
|
|
## print ' Missing jobs are: ' + nSubmit
|
| 755 |
|
|
##
|
| 756 |
|
|
## sys.exit(0)
|
