ViewVC Help

View File | Revision Log | Show Annotations | Root Listing

root/cvsroot/COMP/CRAB/python/cms_cmssw.py

Comparing COMP/CRAB/python/cms_cmssw.py (file contents):
Revision 1.2 by slacapra, Wed Apr 26 15:31:06 2006 UTC vs.
Revision 1.205 by slacapra, Thu Jun 5 16:34:04 2008 UTC

#	Line 2 | Line 2 | from JobType import JobType
2		from crab_logger import Logger
3		from crab_exceptions import *
4		from crab_util import *
5	+	from BlackWhiteListParser import BlackWhiteListParser
6		import common
6	–
7	–	import DataDiscovery
8	–	import DataLocation
7		import Scram
8	+	from LFNBaseName import *
9
10	<	import os, string, re
10	>	import os, string, glob
11
12		class Cmssw(JobType):
13	<	def __init__(self, cfg_params):
13	>	def __init__(self, cfg_params, ncjobs):
14		JobType.__init__(self, 'CMSSW')
15		common.logger.debug(3,'CMSSW::__init__')
16
17	<	self.analisys_common_info = {}
17	>	self.argsList = []
18	>
19	>	self._params = {}
20	>	self.cfg_params = cfg_params
21	>	# init BlackWhiteListParser
22	>	self.blackWhiteListParser = BlackWhiteListParser(cfg_params)
23	>
24	>	self.MaxTarBallSize = float(self.cfg_params.get('EDG.maxtarballsize',9.5))
25	>
26	>	# number of jobs requested to be created, limit obj splitting
27	>	self.ncjobs = ncjobs
28
29		log = common.logger
30	<
30	>
31		self.scram = Scram.Scram(cfg_params)
23	–	scramArea = ''
32		self.additional_inbox_files = []
33		self.scriptExe = ''
34		self.executable = ''
35	+	self.executable_arch = self.scram.getArch()
36		self.tgz_name = 'default.tgz'
37	+	self.scriptName = 'CMSSW.sh'
38	+	self.pset = ''
39	+	self.datasetPath = ''
40	+
41	+	# set FJR file name
42	+	self.fjrFileName = 'crab_fjr.xml'
43
44		self.version = self.scram.getSWVersion()
45	<	common.analisys_common_info['sw_version'] = self.version
45	>	version_array = self.version.split('_')
46	>	self.CMSSW_major = 0
47	>	self.CMSSW_minor = 0
48	>	self.CMSSW_patch = 0
49	>	try:
50	>	self.CMSSW_major = int(version_array[1])
51	>	self.CMSSW_minor = int(version_array[2])
52	>	self.CMSSW_patch = int(version_array[3])
53	>	except:
54	>	msg = "Cannot parse CMSSW version string: " + self.version + " for major and minor release number!"
55	>	raise CrabException(msg)
56
57		### collect Data cards
58	<	try:
59	<	self.owner = cfg_params['CMSSW.owner']
60	<	log.debug(6, "CMSSW::CMSSW(): owner = "+self.owner)
36	<	self.dataset = cfg_params['CMSSW.dataset']
37	<	log.debug(6, "CMSSW::CMSSW(): dataset = "+self.dataset)
38	<	except KeyError:
39	<	msg = "Error: owner and/or dataset not defined "
58	>
59	>	if not cfg_params.has_key('CMSSW.datasetpath'):
60	>	msg = "Error: datasetpath not defined "
61		raise CrabException(msg)
62	+	tmp =  cfg_params['CMSSW.datasetpath']
63	+	log.debug(6, "CMSSW::CMSSW(): datasetPath = "+tmp)
64	+	if string.lower(tmp)=='none':
65	+	self.datasetPath = None
66	+	self.selectNoInput = 1
67	+	else:
68	+	self.datasetPath = tmp
69	+	self.selectNoInput = 0
70
71		self.dataTiers = []
72	<	try:
73	<	tmpDataTiers = string.split(cfg_params['CMSSW.data_tier'],',')
74	<	for tmp in tmpDataTiers:
46	<	tmp=string.strip(tmp)
47	<	self.dataTiers.append(tmp)
48	<	pass
49	<	pass
50	<	except KeyError:
51	<	pass
52	<	log.debug(6, "Cmssw::Cmssw(): dataTiers = "+str(self.dataTiers))
53	<
72	>	self.debugWrap = ''
73	>	self.debug_wrapper = cfg_params.get('USER.debug_wrapper',False)
74	>	if self.debug_wrapper: self.debugWrap='--debug'
75		## now the application
76	<	try:
77	<	self.executable = cfg_params['CMSSW.executable']
57	<	log.debug(6, "CMSSW::CMSSW(): executable = "+self.executable)
58	<	msg = "Default executable cmsRun overridden. Switch to " + self.executable
59	<	log.debug(3,msg)
60	<	except KeyError:
61	<	self.executable = 'cmsRun'
62	<	msg = "User executable not defined. Use cmsRun"
63	<	log.debug(3,msg)
64	<	pass
76	>	self.executable = cfg_params.get('CMSSW.executable','cmsRun')
77	>	log.debug(6, "CMSSW::CMSSW(): executable = "+self.executable)
78
79	<	try:
80	<	self.pset = cfg_params['CMSSW.pset']
81	<	log.debug(6, "Cmssw::Cmssw(): PSet file = "+self.pset)
79	>	if not cfg_params.has_key('CMSSW.pset'):
80	>	raise CrabException("PSet file missing. Cannot run cmsRun ")
81	>	self.pset = cfg_params['CMSSW.pset']
82	>	log.debug(6, "Cmssw::Cmssw(): PSet file = "+self.pset)
83	>	if self.pset.lower() != 'none' :
84		if (not os.path.exists(self.pset)):
85		raise CrabException("User defined PSet file "+self.pset+" does not exist")
86	<	except KeyError:
87	<	raise CrabException("PSet file missing. Cannot run cmsRun ")
86	>	else:
87	>	self.pset = None
88
89		# output files
90	<	try:
91	<	self.output_file = []
90	>	## stuff which must be returned always via sandbox
91	>	self.output_file_sandbox = []
92
93	<	tmp = cfg_params['CMSSW.output_file']
94	<	if tmp != '':
80	<	tmpOutFiles = string.split(cfg_params['CMSSW.output_file'],',')
81	<	log.debug(7, 'cmssw::cmssw(): output files '+str(tmpOutFiles))
82	<	for tmp in tmpOutFiles:
83	<	tmp=string.strip(tmp)
84	<	self.output_file.append(tmp)
85	<	pass
93	>	# add fjr report by default via sandbox
94	>	self.output_file_sandbox.append(self.fjrFileName)
95
96	<	else:
97	<	log.message("No output file defined: only stdout/err will be available")
96	>	# other output files to be returned via sandbox or copied to SE
97	>	self.output_file = []
98	>	tmp = cfg_params.get('CMSSW.output_file',None)
99	>	if tmp :
100	>	tmpOutFiles = string.split(tmp,',')
101	>	log.debug(7, 'cmssw::cmssw(): output files '+str(tmpOutFiles))
102	>	for tmp in tmpOutFiles:
103	>	tmp=string.strip(tmp)
104	>	self.output_file.append(tmp)
105		pass
106	<	pass
107	<	except KeyError:
108	<	log.message("No output file defined: only stdout/err will be available")
93	<	pass
106	>	else:
107	>	log.message("No output file defined: only stdout/err and the CRAB Framework Job Report will be available\n")
108	>	pass
109
110		# script_exe file as additional file in inputSandbox
111	<	try:
112	<	self.scriptExe = cfg_params['CMSSW.script_exe']
113	<	self.additional_inbox_files.append(self.scriptExe)
114	<	except KeyError:
115	<	pass
116	<	if self.scriptExe != '':
117	<	if os.path.isfile(self.scriptExe):
118	<	pass
119	<	else:
120	<	log.message("WARNING. file "+self.scriptExe+" not found")
121	<	sys.exit()
122	<
111	>	self.scriptExe = cfg_params.get('USER.script_exe',None)
112	>	if self.scriptExe :
113	>	if not os.path.isfile(self.scriptExe):
114	>	msg ="ERROR. file "+self.scriptExe+" not found"
115	>	raise CrabException(msg)
116	>	self.additional_inbox_files.append(string.strip(self.scriptExe))
117	>
118	>	if self.datasetPath == None and self.pset == None and self.scriptExe == '' :
119	>	msg ="Error. script_exe  not defined"
120	>	raise CrabException(msg)
121	>
122	>	# use parent files...
123	>	self.useParent = self.cfg_params.get('CMSSW.use_parent',False)
124	>
125		## additional input files
126	<	try:
127	<	tmpAddFiles = string.split(cfg_params['CMSSW.additional_input_files'],',')
126	>	if cfg_params.has_key('USER.additional_input_files'):
127	>	tmpAddFiles = string.split(cfg_params['USER.additional_input_files'],',')
128		for tmp in tmpAddFiles:
129	<	tmp=string.strip(tmp)
130	<	self.additional_inbox_files.append(tmp)
129	>	tmp = string.strip(tmp)
130	>	dirname = ''
131	>	if not tmp[0]=="/": dirname = "."
132	>	files = []
133	>	if string.find(tmp,"*")>-1:
134	>	files = glob.glob(os.path.join(dirname, tmp))
135	>	if len(files)==0:
136	>	raise CrabException("No additional input file found with this pattern: "+tmp)
137	>	else:
138	>	files.append(tmp)
139	>	for file in files:
140	>	if not os.path.exists(file):
141	>	raise CrabException("Additional input file not found: "+file)
142	>	pass
143	>	self.additional_inbox_files.append(string.strip(file))
144		pass
145		pass
146	<	except KeyError:
147	<	pass
146	>	common.logger.debug(5,"Additional input files: "+str(self.additional_inbox_files))
147	>	pass
148
149	<	try:
150	<	self.total_number_of_events = int(cfg_params['CMSSW.total_number_of_events'])
151	<	except KeyError:
152	<	msg = 'Must define total_number_of_events and job_number_of_events'
153	<	raise CrabException(msg)
154	<
155	<	#Marco: FirstEvent is nolonger used inside PSet
126	<	#        try:
127	<	#            self.first = int(cfg_params['CMSSW.first_event'])
128	<	#        except KeyError:
129	<	#            self.first = 0
130	<	#            pass
131	<	#        log.debug(6, "Orca::Orca(): total number of events = "+`self.total_number_of_events`)
132	<	#log.debug(6, "Orca::Orca(): events per job = "+`self.job_number_of_events`)
133	<	#        log.debug(6, "Orca::Orca(): first event = "+`self.first`)
134	<
135	<	CEBlackList = []
136	<	try:
137	<	tmpBad = string.split(cfg_params['EDG.ce_black_list'],',')
138	<	for tmp in tmpBad:
139	<	tmp=string.strip(tmp)
140	<	CEBlackList.append(tmp)
141	<	except KeyError:
142	<	pass
149	>	## Events per job
150	>	if cfg_params.has_key('CMSSW.events_per_job'):
151	>	self.eventsPerJob =int( cfg_params['CMSSW.events_per_job'])
152	>	self.selectEventsPerJob = 1
153	>	else:
154	>	self.eventsPerJob = -1
155	>	self.selectEventsPerJob = 0
156
157	<	self.reCEBlackList=[]
158	<	for bad in CEBlackList:
159	<	self.reCEBlackList.append(re.compile( bad ))
157	>	## number of jobs
158	>	if cfg_params.has_key('CMSSW.number_of_jobs'):
159	>	self.theNumberOfJobs =int( cfg_params['CMSSW.number_of_jobs'])
160	>	self.selectNumberOfJobs = 1
161	>	else:
162	>	self.theNumberOfJobs = 0
163	>	self.selectNumberOfJobs = 0
164
165	<	common.logger.debug(5,'CEBlackList: '+str(CEBlackList))
165	>	if cfg_params.has_key('CMSSW.total_number_of_events'):
166	>	self.total_number_of_events = int(cfg_params['CMSSW.total_number_of_events'])
167	>	self.selectTotalNumberEvents = 1
168	>	if self.selectNumberOfJobs  == 1:
169	>	if (self.total_number_of_events != -1) and int(self.total_number_of_events) < int(self.theNumberOfJobs):
170	>	msg = 'Must specify at least one event per job. total_number_of_events > number_of_jobs '
171	>	raise CrabException(msg)
172	>	else:
173	>	self.total_number_of_events = 0
174	>	self.selectTotalNumberEvents = 0
175
176	<	CEWhiteList = []
177	<	try:
178	<	tmpGood = string.split(cfg_params['EDG.ce_white_list'],',')
179	<	#tmpGood = ['cern']
180	<	for tmp in tmpGood:
181	<	tmp=string.strip(tmp)
182	<	#if (tmp == 'cnaf'): tmp = 'webserver' ########## warning: temp. patch
183	<	CEWhiteList.append(tmp)
184	<	except KeyError:
185	<	pass
176	>	if self.pset != None:
177	>	if ( (self.selectTotalNumberEvents + self.selectEventsPerJob + self.selectNumberOfJobs) != 2 ):
178	>	msg = 'Must define exactly two of total_number_of_events, events_per_job, or number_of_jobs.'
179	>	raise CrabException(msg)
180	>	else:
181	>	if (self.selectNumberOfJobs == 0):
182	>	msg = 'Must specify  number_of_jobs.'
183	>	raise CrabException(msg)
184	>
185	>	## New method of dealing with seeds
186	>	self.incrementSeeds = []
187	>	self.preserveSeeds = []
188	>	if cfg_params.has_key('CMSSW.preserve_seeds'):
189	>	tmpList = cfg_params['CMSSW.preserve_seeds'].split(',')
190	>	for tmp in tmpList:
191	>	tmp.strip()
192	>	self.preserveSeeds.append(tmp)
193	>	if cfg_params.has_key('CMSSW.increment_seeds'):
194	>	tmpList = cfg_params['CMSSW.increment_seeds'].split(',')
195	>	for tmp in tmpList:
196	>	tmp.strip()
197	>	self.incrementSeeds.append(tmp)
198	>
199	>	## Old method of dealing with seeds
200	>	## FUTURE: This is for old CMSSW and old CRAB. Can throw exceptions after a couple of CRAB releases and then
201	>	## remove
202	>	self.sourceSeed = cfg_params.get('CMSSW.pythia_seed',None)
203	>	if self.sourceSeed:
204	>	print "pythia_seed is a deprecated parameter. Use preserve_seeds or increment_seeds in the future.\n","Added to increment_seeds."
205	>	self.incrementSeeds.append('sourceSeed')
206	>	self.incrementSeeds.append('theSource')
207	>
208	>	self.sourceSeedVtx = cfg_params.get('CMSSW.vtx_seed',None)
209	>	if self.sourceSeedVtx:
210	>	print "vtx_seed is a deprecated parameter. Use preserve_seeds or increment_seeds in the future.\n","Added to increment_seeds."
211	>	self.incrementSeeds.append('VtxSmeared')
212	>
213	>	self.sourceSeedG4 = cfg_params.get('CMSSW.g4_seed',None)
214	>	if self.sourceSeedG4:
215	>	print "g4_seed is a deprecated parameter. Use preserve_seeds or increment_seeds in the future.\n","Added to increment_seeds."
216	>	self.incrementSeeds.append('g4SimHits')
217	>
218	>	self.sourceSeedMix = cfg_params.get('CMSSW.mix_seed',None)
219	>	if self.sourceSeedMix:
220	>	print "mix_seed is a deprecated parameter. Use preserve_seeds or increment_seeds in the future.\n","Added to increment_seeds."
221	>	self.incrementSeeds.append('mix')
222	>
223	>	self.firstRun = cfg_params.get('CMSSW.first_run',None)
224	>
225	>	if self.pset != None: #CarlosDaniele
226	>	import PsetManipulator as pp
227	>	PsetEdit = pp.PsetManipulator(self.pset) #Daniele Pset
228
229	<	#print 'CEWhiteList: ',CEWhiteList
162	<	self.reCEWhiteList=[]
163	<	for Good in CEWhiteList:
164	<	self.reCEWhiteList.append(re.compile( Good ))
229	>	# Copy/return
230
231	<	common.logger.debug(5,'CEWhiteList: '+str(CEWhiteList))
231	>	self.copy_data = int(cfg_params.get('USER.copy_data',0))
232	>	self.return_data = int(cfg_params.get('USER.return_data',0))
233
234		#DBSDLS-start
235	<	## Initialize the variables that are extracted from DBS/DLS and needed in other places of the code
235	>	## Initialize the variables that are extracted from DBS/DLS and needed in other places of the code
236		self.maxEvents=0  # max events available   ( --> check the requested nb. of evts in Creator.py)
237		self.DBSPaths={}  # all dbs paths requested ( --> input to the site local discovery script)
238	+	self.jobDestination=[]  # Site destination(s) for each job (list of lists)
239		## Perform the data location and discovery (based on DBS/DLS)
240	<	self.DataDiscoveryAndLocation(cfg_params)
241	<	#DBSDLS-end
240	>	## SL: Don't if NONE is specified as input (pythia use case)
241	>	blockSites = {}
242	>	if self.datasetPath:
243	>	blockSites = self.DataDiscoveryAndLocation(cfg_params)
244	>	#DBSDLS-end
245	>
246
247	+	## Select Splitting
248	+	if self.selectNoInput:
249	+	if self.pset == None:
250	+	self.jobSplittingForScript()
251	+	else:
252	+	self.jobSplittingNoInput()
253	+	else:
254	+	self.jobSplittingByBlocks(blockSites)
255	+
256	+	# modify Pset
257	+	if self.pset != None:
258	+	try:
259	+	# Add FrameworkJobReport to parameter-set, set max events.
260	+	# Reset later for data jobs by writeCFG which does all modifications
261	+	PsetEdit.addCrabFJR(self.fjrFileName) # FUTURE: Job report addition not needed by CMSSW>1.5
262	+	PsetEdit.maxEvent(self.eventsPerJob)
263	+	PsetEdit.psetWriter(self.configFilename())
264	+	except:
265	+	msg='Error while manipulating ParameterSet: exiting...'
266	+	raise CrabException(msg)
267		self.tgzNameWithPath = self.getTarBall(self.executable)
268
269		def DataDiscoveryAndLocation(self, cfg_params):
270
271	<	fun = "CMSSW::DataDiscoveryAndLocation()"
271	>	import DataDiscovery
272	>	import DataLocation
273	>	common.logger.debug(10,"CMSSW::DataDiscoveryAndLocation()")
274	>
275	>	datasetPath=self.datasetPath
276
277		## Contact the DBS
278	+	common.logger.message("Contacting Data Discovery Services ...")
279		try:
280	<	self.pubdata=DataDiscovery.DataDiscovery(self.owner,
185	<	self.dataset,
186	<	self.dataTiers,
187	<	cfg_params)
280	>	self.pubdata=DataDiscovery.DataDiscovery(datasetPath, cfg_params)
281		self.pubdata.fetchDBSInfo()
282
283		except DataDiscovery.NotExistingDatasetError, ex :
284		msg = 'ERROR ***: failed Data Discovery in DBS : %s'%ex.getErrorMessage()
285		raise CrabException(msg)
193	–
286		except DataDiscovery.NoDataTierinProvenanceError, ex :
287		msg = 'ERROR ***: failed Data Discovery in DBS : %s'%ex.getErrorMessage()
288		raise CrabException(msg)
289		except DataDiscovery.DataDiscoveryError, ex:
290	<	msg = 'ERROR ***: failed Data Discovery in DBS  %s'%ex.getErrorMessage()
290	>	msg = 'ERROR ***: failed Data Discovery in DBS :  %s'%ex.getErrorMessage()
291		raise CrabException(msg)
292
293	<	## get list of all required data in the form of dbs paths  (dbs path = /dataset/datatier/owner)
294	<	self.DBSPaths=self.pubdata.getDBSPaths()
295	<	common.logger.message("Required data are : ")
296	<	for path in self.DBSPaths:
205	<	common.logger.message(" --> "+path )
293	>	self.filesbyblock=self.pubdata.getFiles()
294	>	self.eventsbyblock=self.pubdata.getEventsPerBlock()
295	>	self.eventsbyfile=self.pubdata.getEventsPerFile()
296	>	self.parentFiles=self.pubdata.getParent()
297
298		## get max number of events
299	<	common.logger.debug(10,"number of events for primary fileblocks %i"%self.pubdata.getMaxEvents())
209	<	self.maxEvents=self.pubdata.getMaxEvents() ##  self.maxEvents used in Creator.py
210	<	common.logger.message("\nThe number of available events is %s"%self.maxEvents)
211	<
212	<	## get fileblocks corresponding to the required data
213	<	fb=self.pubdata.getFileBlocks()
214	<	common.logger.debug(5,"fileblocks are %s"%fb)
299	>	self.maxEvents=self.pubdata.getMaxEvents()
300
301		## Contact the DLS and build a list of sites hosting the fileblocks
302		try:
303	<	dataloc=DataLocation.DataLocation(self.pubdata.getFileBlocks(),cfg_params)
303	>	dataloc=DataLocation.DataLocation(self.filesbyblock.keys(),cfg_params)
304		dataloc.fetchDLSInfo()
305		except DataLocation.DataLocationError , ex:
306		msg = 'ERROR ***: failed Data Location in DLS \n %s '%ex.getErrorMessage()
307		raise CrabException(msg)
223	–
224	–	allsites=dataloc.getSites()
225	–	common.logger.debug(5,"sites are %s"%allsites)
226	–	sites=self.checkBlackList(allsites)
227	–	common.logger.debug(5,"sites are (after black list) %s"%sites)
228	–	sites=self.checkWhiteList(sites)
229	–	common.logger.debug(5,"sites are (after white list) %s"%sites)
308
309	<	if len(sites)==0:
310	<	msg = 'No sites hosting all the needed data! Exiting... '
311	<	raise CrabException(msg)
312	<	common.logger.message("List of Sites hosting the data : "+str(sites))
313	<	common.logger.debug(6, "List of Sites: "+str(sites))
314	<	common.analisys_common_info['sites']=sites    ## used in SchedulerEdg.py in createSchScript
315	<	return
316	<
317	<	def checkBlackList(self, allSites):
318	<	if len(self.reCEBlackList)==0: return allSites
319	<	sites = []
320	<	for site in allSites:
243	<	common.logger.debug(10,'Site '+site)
244	<	good=1
245	<	for re in self.reCEBlackList:
246	<	if re.search(site):
247	<	common.logger.message('CE in black list, skipping site '+site)
248	<	good=0
249	<	pass
250	<	if good: sites.append(site)
251	<	if len(sites) == 0:
252	<	common.logger.debug(3,"No sites found after BlackList")
309	>
310	>	sites = dataloc.getSites()
311	>	allSites = []
312	>	listSites = sites.values()
313	>	for listSite in listSites:
314	>	for oneSite in listSite:
315	>	allSites.append(oneSite)
316	>	allSites = self.uniquelist(allSites)
317	>
318	>	# screen output
319	>	common.logger.message("Requested dataset: " + datasetPath + " has " + str(self.maxEvents) + " events in " + str(len(self.filesbyblock.keys())) + " blocks.\n")
320	>
321		return sites
322
323	<	def checkWhiteList(self, allsites):
323	>	def jobSplittingByBlocks(self, blockSites):
324	>	"""
325	>	Perform job splitting. Jobs run over an integer number of files
326	>	and no more than one block.
327	>	ARGUMENT: blockSites: dictionary with blocks as keys and list of host sites as values
328	>	REQUIRES: self.selectTotalNumberEvents, self.selectEventsPerJob, self.selectNumberofJobs,
329	>	self.total_number_of_events, self.eventsPerJob, self.theNumberOfJobs,
330	>	self.maxEvents, self.filesbyblock
331	>	SETS: self.jobDestination - Site destination(s) for each job (a list of lists)
332	>	self.total_number_of_jobs - Total # of jobs
333	>	self.list_of_args - File(s) job will run on (a list of lists)
334	>	"""
335
336	<	if len(self.reCEWhiteList)==0: return pubDBUrls
337	<	sites = []
338	<	for site in allsites:
339	<	#print 'connecting to the URL ',url
340	<	good=0
341	<	for re in self.reCEWhiteList:
342	<	if re.search(site):
343	<	common.logger.debug(5,'CE in white list, adding site '+site)
344	<	good=1
345	<	if not good: continue
346	<	sites.append(site)
347	<	if len(sites) == 0:
348	<	common.logger.message("No sites found after WhiteList\n")
336	>	# ---- Handle the possible job splitting configurations ---- #
337	>	if (self.selectTotalNumberEvents):
338	>	totalEventsRequested = self.total_number_of_events
339	>	if (self.selectEventsPerJob):
340	>	eventsPerJobRequested = self.eventsPerJob
341	>	if (self.selectNumberOfJobs):
342	>	totalEventsRequested = self.theNumberOfJobs * self.eventsPerJob
343	>
344	>	# If user requested all the events in the dataset
345	>	if (totalEventsRequested == -1):
346	>	eventsRemaining=self.maxEvents
347	>	# If user requested more events than are in the dataset
348	>	elif (totalEventsRequested > self.maxEvents):
349	>	eventsRemaining = self.maxEvents
350	>	common.logger.message("Requested "+str(self.total_number_of_events)+ " events, but only "+str(self.maxEvents)+" events are available.")
351	>	# If user requested less events than are in the dataset
352		else:
353	<	common.logger.debug(5,"Selected sites via WhiteList are "+str(sites)+"\n")
354	<	return sites
353	>	eventsRemaining = totalEventsRequested
354	>
355	>	# If user requested more events per job than are in the dataset
356	>	if (self.selectEventsPerJob and eventsPerJobRequested > self.maxEvents):
357	>	eventsPerJobRequested = self.maxEvents
358	>
359	>	# For user info at end
360	>	totalEventCount = 0
361	>
362	>	if (self.selectTotalNumberEvents and self.selectNumberOfJobs):
363	>	eventsPerJobRequested = int(eventsRemaining/self.theNumberOfJobs)
364	>
365	>	if (self.selectNumberOfJobs):
366	>	common.logger.message("May not create the exact number_of_jobs requested.")
367	>
368	>	if ( self.ncjobs == 'all' ) :
369	>	totalNumberOfJobs = 999999999
370	>	else :
371	>	totalNumberOfJobs = self.ncjobs
372	>
373	>	blocks = blockSites.keys()
374	>	blockCount = 0
375	>	# Backup variable in case self.maxEvents counted events in a non-included block
376	>	numBlocksInDataset = len(blocks)
377	>
378	>	jobCount = 0
379	>	list_of_lists = []
380	>
381	>	# list tracking which jobs are in which jobs belong to which block
382	>	jobsOfBlock = {}
383	>
384	>	# ---- Iterate over the blocks in the dataset until ---- #
385	>	# ---- we've met the requested total # of events    ---- #
386	>	while ( (eventsRemaining > 0) and (blockCount < numBlocksInDataset) and (jobCount < totalNumberOfJobs)):
387	>	block = blocks[blockCount]
388	>	blockCount += 1
389	>	if block not in jobsOfBlock.keys() :
390	>	jobsOfBlock[block] = []
391	>
392	>	if self.eventsbyblock.has_key(block) :
393	>	numEventsInBlock = self.eventsbyblock[block]
394	>	common.logger.debug(5,'Events in Block File '+str(numEventsInBlock))
395	>
396	>	files = self.filesbyblock[block]
397	>	numFilesInBlock = len(files)
398	>	if (numFilesInBlock <= 0):
399	>	continue
400	>	fileCount = 0
401	>
402	>	# ---- New block => New job ---- #
403	>	parString = ""
404	>	# counter for number of events in files currently worked on
405	>	filesEventCount = 0
406	>	# flag if next while loop should touch new file
407	>	newFile = 1
408	>	# job event counter
409	>	jobSkipEventCount = 0
410	>
411	>	# ---- Iterate over the files in the block until we've met the requested ---- #
412	>	# ---- total # of events or we've gone over all the files in this block  ---- #
413	>	pString=''
414	>	while ( (eventsRemaining > 0) and (fileCount < numFilesInBlock) and (jobCount < totalNumberOfJobs) ):
415	>	file = files[fileCount]
416	>	if self.useParent:
417	>	parent = self.parentFiles[file]
418	>	for f in parent :
419	>	pString += '\\\"' + f + '\\\"\,'
420	>	common.logger.debug(6, "File "+str(file)+" has the following parents: "+str(parent))
421	>	common.logger.write("File "+str(file)+" has the following parents: "+str(parent))
422	>	if newFile :
423	>	try:
424	>	numEventsInFile = self.eventsbyfile[file]
425	>	common.logger.debug(6, "File "+str(file)+" has "+str(numEventsInFile)+" events")
426	>	# increase filesEventCount
427	>	filesEventCount += numEventsInFile
428	>	# Add file to current job
429	>	parString += '\\\"' + file + '\\\"\,'
430	>	newFile = 0
431	>	except KeyError:
432	>	common.logger.message("File "+str(file)+" has unknown number of events: skipping")
433	>
434	>	eventsPerJobRequested = min(eventsPerJobRequested, eventsRemaining)
435	>	# if less events in file remain than eventsPerJobRequested
436	>	if ( filesEventCount - jobSkipEventCount < eventsPerJobRequested):
437	>	# if last file in block
438	>	if ( fileCount == numFilesInBlock-1 ) :
439	>	# end job using last file, use remaining events in block
440	>	# close job and touch new file
441	>	fullString = parString[:-2]
442	>	if self.useParent:
443	>	fullParentString = pString[:-2]
444	>	list_of_lists.append([fullString,fullParentString,str(-1),str(jobSkipEventCount)])
445	>	else:
446	>	list_of_lists.append([fullString,str(-1),str(jobSkipEventCount)])
447	>	common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(filesEventCount - jobSkipEventCount)+" events (last file in block).")
448	>	self.jobDestination.append(blockSites[block])
449	>	common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(self.jobDestination[jobCount]))
450	>	# fill jobs of block dictionary
451	>	jobsOfBlock[block].append(jobCount+1)
452	>	# reset counter
453	>	jobCount = jobCount + 1
454	>	totalEventCount = totalEventCount + filesEventCount - jobSkipEventCount
455	>	eventsRemaining = eventsRemaining - filesEventCount + jobSkipEventCount
456	>	jobSkipEventCount = 0
457	>	# reset file
458	>	pString = ""
459	>	parString = ""
460	>	filesEventCount = 0
461	>	newFile = 1
462	>	fileCount += 1
463	>	else :
464	>	# go to next file
465	>	newFile = 1
466	>	fileCount += 1
467	>	# if events in file equal to eventsPerJobRequested
468	>	elif ( filesEventCount - jobSkipEventCount == eventsPerJobRequested ) :
469	>	# close job and touch new file
470	>	fullString = parString[:-2]
471	>	if self.useParent:
472	>	fullParentString = pString[:-2]
473	>	list_of_lists.append([fullString,fullParentString,str(eventsPerJobRequested),str(jobSkipEventCount)])
474	>	else:
475	>	list_of_lists.append([fullString,str(eventsPerJobRequested),str(jobSkipEventCount)])
476	>	common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(eventsPerJobRequested)+" events.")
477	>	self.jobDestination.append(blockSites[block])
478	>	common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(self.jobDestination[jobCount]))
479	>	jobsOfBlock[block].append(jobCount+1)
480	>	# reset counter
481	>	jobCount = jobCount + 1
482	>	totalEventCount = totalEventCount + eventsPerJobRequested
483	>	eventsRemaining = eventsRemaining - eventsPerJobRequested
484	>	jobSkipEventCount = 0
485	>	# reset file
486	>	pString = ""
487	>	parString = ""
488	>	filesEventCount = 0
489	>	newFile = 1
490	>	fileCount += 1
491	>
492	>	# if more events in file remain than eventsPerJobRequested
493	>	else :
494	>	# close job but don't touch new file
495	>	fullString = parString[:-2]
496	>	if self.useParent:
497	>	fullParentString = pString[:-2]
498	>	list_of_lists.append([fullString,fullParentString,str(eventsPerJobRequested),str(jobSkipEventCount)])
499	>	else:
500	>	list_of_lists.append([fullString,str(eventsPerJobRequested),str(jobSkipEventCount)])
501	>	common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(eventsPerJobRequested)+" events.")
502	>	self.jobDestination.append(blockSites[block])
503	>	common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(self.jobDestination[jobCount]))
504	>	jobsOfBlock[block].append(jobCount+1)
505	>	# increase counter
506	>	jobCount = jobCount + 1
507	>	totalEventCount = totalEventCount + eventsPerJobRequested
508	>	eventsRemaining = eventsRemaining - eventsPerJobRequested
509	>	# calculate skip events for last file
510	>	# use filesEventCount (contains several files), jobSkipEventCount and eventsPerJobRequest
511	>	jobSkipEventCount = eventsPerJobRequested - (filesEventCount - jobSkipEventCount - self.eventsbyfile[file])
512	>	# remove all but the last file
513	>	filesEventCount = self.eventsbyfile[file]
514	>	if self.useParent:
515	>	for f in parent : pString += '\\\"' + f + '\\\"\,'
516	>	parString = '\\\"' + file + '\\\"\,'
517	>	pass # END if
518	>	pass # END while (iterate over files in the block)
519	>	pass # END while (iterate over blocks in the dataset)
520	>	self.ncjobs = self.total_number_of_jobs = jobCount
521	>	if (eventsRemaining > 0 and jobCount < totalNumberOfJobs ):
522	>	common.logger.message("Could not run on all requested events because some blocks not hosted at allowed sites.")
523	>	common.logger.message(str(jobCount)+" job(s) can run on "+str(totalEventCount)+" events.\n")
524	>
525	>	# screen output
526	>	screenOutput = "List of jobs and available destination sites:\n\n"
527	>
528	>	# keep trace of block with no sites to print a warning at the end
529	>	noSiteBlock = []
530	>	bloskNoSite = []
531	>
532	>	blockCounter = 0
533	>	for block in blocks:
534	>	if block in jobsOfBlock.keys() :
535	>	blockCounter += 1
536	>	screenOutput += "Block %5i: jobs %20s: sites: %s\n" % (blockCounter,spanRanges(jobsOfBlock[block]),
537	>	','.join(self.blackWhiteListParser.checkWhiteList(self.blackWhiteListParser.checkBlackList(blockSites[block],block),block)))
538	>	if len(self.blackWhiteListParser.checkWhiteList(self.blackWhiteListParser.checkBlackList(blockSites[block],block),block)) == 0:
539	>	noSiteBlock.append( spanRanges(jobsOfBlock[block]) )
540	>	bloskNoSite.append( blockCounter )
541	>
542	>	common.logger.message(screenOutput)
543	>	if len(noSiteBlock) > 0 and len(bloskNoSite) > 0:
544	>	msg = 'WARNING: No sites are hosting any part of data for block:\n                '
545	>	virgola = ""
546	>	if len(bloskNoSite) > 1:
547	>	virgola = ","
548	>	for block in bloskNoSite:
549	>	msg += ' ' + str(block) + virgola
550	>	msg += '\n               Related jobs:\n                 '
551	>	virgola = ""
552	>	if len(noSiteBlock) > 1:
553	>	virgola = ","
554	>	for range_jobs in noSiteBlock:
555	>	msg += str(range_jobs) + virgola
556	>	msg += '\n               will not be submitted and this block of data can not be analyzed!\n'
557	>	if self.cfg_params.has_key('EDG.se_white_list'):
558	>	msg += 'WARNING: SE White List: '+self.cfg_params['EDG.se_white_list']+'\n'
559	>	msg += '(Hint: By whitelisting you force the job to run at this particular site(s).\n'
560	>	msg += 'Please check if the dataset is available at this site!)\n'
561	>	if self.cfg_params.has_key('EDG.ce_white_list'):
562	>	msg += 'WARNING: CE White List: '+self.cfg_params['EDG.ce_white_list']+'\n'
563	>	msg += '(Hint: By whitelisting you force the job to run at this particular site(s).\n'
564	>	msg += 'Please check if the dataset is available at this site!)\n'
565	>
566	>	common.logger.message(msg)
567	>
568	>	self.list_of_args = list_of_lists
569	>	return
570	>
571	>	def jobSplittingNoInput(self):
572	>	"""
573	>	Perform job splitting based on number of event per job
574	>	"""
575	>	common.logger.debug(5,'Splitting per events')
576	>
577	>	if (self.selectEventsPerJob):
578	>	common.logger.message('Required '+str(self.eventsPerJob)+' events per job ')
579	>	if (self.selectNumberOfJobs):
580	>	common.logger.message('Required '+str(self.theNumberOfJobs)+' jobs in total ')
581	>	if (self.selectTotalNumberEvents):
582	>	common.logger.message('Required '+str(self.total_number_of_events)+' events in total ')
583	>
584	>	if (self.total_number_of_events < 0):
585	>	msg='Cannot split jobs per Events with "-1" as total number of events'
586	>	raise CrabException(msg)
587	>
588	>	if (self.selectEventsPerJob):
589	>	if (self.selectTotalNumberEvents):
590	>	self.total_number_of_jobs = int(self.total_number_of_events/self.eventsPerJob)
591	>	elif(self.selectNumberOfJobs) :
592	>	self.total_number_of_jobs =self.theNumberOfJobs
593	>	self.total_number_of_events =int(self.theNumberOfJobs*self.eventsPerJob)
594	>
595	>	elif (self.selectNumberOfJobs) :
596	>	self.total_number_of_jobs = self.theNumberOfJobs
597	>	self.eventsPerJob = int(self.total_number_of_events/self.total_number_of_jobs)
598	>
599	>	common.logger.debug(5,'N jobs  '+str(self.total_number_of_jobs))
600	>
601	>	# is there any remainder?
602	>	check = int(self.total_number_of_events) - (int(self.total_number_of_jobs)*self.eventsPerJob)
603	>
604	>	common.logger.debug(5,'Check  '+str(check))
605	>
606	>	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')
607	>	if check > 0:
608	>	common.logger.message('Warning: asked '+str(self.total_number_of_events)+' but can do only '+str(int(self.total_number_of_jobs)*self.eventsPerJob))
609	>
610	>	# argument is seed number.$i
611	>	self.list_of_args = []
612	>	for i in range(self.total_number_of_jobs):
613	>	## Since there is no input, any site is good
614	>	self.jobDestination.append([""]) #must be empty to write correctly the xml
615	>	args=[]
616	>	if (self.firstRun):
617	>	## pythia first run
618	>	args.append(str(self.firstRun)+str(i))
619	>	self.list_of_args.append(args)
620	>
621	>	return
622	>
623	>
624	>	def jobSplittingForScript(self):
625	>	"""
626	>	Perform job splitting based on number of job
627	>	"""
628	>	common.logger.debug(5,'Splitting per job')
629	>	common.logger.message('Required '+str(self.theNumberOfJobs)+' jobs in total ')
630	>
631	>	self.total_number_of_jobs = self.theNumberOfJobs
632	>
633	>	common.logger.debug(5,'N jobs  '+str(self.total_number_of_jobs))
634	>
635	>	common.logger.message(str(self.total_number_of_jobs)+' jobs can be created')
636	>
637	>	# argument is seed number.$i
638	>	self.list_of_args = []
639	>	for i in range(self.total_number_of_jobs):
640	>	self.jobDestination.append([""])
641	>	self.list_of_args.append([str(i)])
642	>	return
643	>
644	>	def split(self, jobParams):
645	>
646	>	njobs = self.total_number_of_jobs
647	>	arglist = self.list_of_args
648	>	# create the empty structure
649	>	for i in range(njobs):
650	>	jobParams.append("")
651	>
652	>	listID=[]
653	>	listField=[]
654	>	for job in range(njobs):
655	>	jobParams[job] = arglist[job]
656	>	listID.append(job+1)
657	>	job_ToSave ={}
658	>	concString = ' '
659	>	argu=''
660	>	if len(jobParams[job]):
661	>	argu +=   concString.join(jobParams[job] )
662	>	job_ToSave['arguments']= str(job+1)+' '+argu
663	>	job_ToSave['dlsDestination']= self.jobDestination[job]
664	>	listField.append(job_ToSave)
665	>	msg="Job "+str(job)+" Arguments:   "+str(job+1)+" "+argu+"\n"  \
666	>	+"                     Destination: "+str(self.jobDestination[job])
667	>	common.logger.debug(5,msg)
668	>	common._db.updateJob_(listID,listField)
669	>	self.argsList = (len(jobParams[0])+1)
670	>
671	>	return
672	>
673	>	def numberOfJobs(self):
674	>	return self.total_number_of_jobs
675
676		def getTarBall(self, exe):
677		"""
678		Return the TarBall with lib and exe
679		"""
680	<
279	<	# if it exist, just return it
280	<	self.tgzNameWithPath = common.work_space.shareDir()+self.tgz_name
680	>	self.tgzNameWithPath = common.work_space.pathForTgz()+'share/'+self.tgz_name
681		if os.path.exists(self.tgzNameWithPath):
682		return self.tgzNameWithPath
683
#	Line 290 | Line 690 | class Cmssw(JobType):
690
691		# First of all declare the user Scram area
692		swArea = self.scram.getSWArea_()
293	–	#print "swArea = ", swArea
294	–	swVersion = self.scram.getSWVersion()
295	–	#print "swVersion = ", swVersion
693		swReleaseTop = self.scram.getReleaseTop_()
694	<	#print "swReleaseTop = ", swReleaseTop
298	<
694	>
695		## check if working area is release top
696		if swReleaseTop == '' or swArea == swReleaseTop:
697	+	common.logger.debug(3,"swArea = "+swArea+" swReleaseTop ="+swReleaseTop)
698		return
699
700	<	filesToBeTarred = []
701	<	## First find the executable
702	<	if (self.executable != ''):
703	<	exeWithPath = self.scram.findFile_(executable)
704	<	#           print exeWithPath
705	<	if ( not exeWithPath ):
706	<	raise CrabException('User executable '+executable+' not found')
707	<
708	<	## then check if it's private or not
709	<	if exeWithPath.find(swReleaseTop) == -1:
710	<	# the exe is private, so we must ship
711	<	common.logger.debug(5,"Exe "+exeWithPath+" to be tarred")
712	<	path = swArea+'/'
713	<	exe = string.replace(exeWithPath, path,'')
714	<	filesToBeTarred.append(exe)
715	<	pass
716	<	else:
717	<	# the exe is from release, we'll find it on WN
718	<	pass
719	<
720	<	## Now get the libraries: only those in local working area
721	<	libDir = 'lib'
722	<	lib = swArea+'/' +libDir
723	<	common.logger.debug(5,"lib "+lib+" to be tarred")
724	<	if os.path.exists(lib):
725	<	filesToBeTarred.append(libDir)
726	<
727	<	## Now check if the Data dir is present
728	<	dataDir = 'src/Data/'
729	<	if os.path.isdir(swArea+'/'+dataDir):
730	<	filesToBeTarred.append(dataDir)
731	<
732	<	## Create the tar-ball
733	<	if len(filesToBeTarred)>0:
734	<	cwd = os.getcwd()
735	<	os.chdir(swArea)
736	<	tarcmd = 'tar zcvf ' + self.tgzNameWithPath + ' '
737	<	for line in filesToBeTarred:
738	<	tarcmd = tarcmd + line + ' '
739	<	cout = runCommand(tarcmd)
740	<	if not cout:
741	<	raise CrabException('Could not create tar-ball')
742	<	os.chdir(cwd)
743	<	else:
744	<	common.logger.debug(5,"No files to be to be tarred")
745	<
746	<	return
747	<
748	<	def wsSetupEnvironment(self, nj):
700	>	import tarfile
701	>	try: # create tar ball
702	>	tar = tarfile.open(self.tgzNameWithPath, "w:gz")
703	>	## First find the executable
704	>	if (self.executable != ''):
705	>	exeWithPath = self.scram.findFile_(executable)
706	>	if ( not exeWithPath ):
707	>	raise CrabException('User executable '+executable+' not found')
708	>
709	>	## then check if it's private or not
710	>	if exeWithPath.find(swReleaseTop) == -1:
711	>	# the exe is private, so we must ship
712	>	common.logger.debug(5,"Exe "+exeWithPath+" to be tarred")
713	>	path = swArea+'/'
714	>	# distinguish case when script is in user project area or given by full path somewhere else
715	>	if exeWithPath.find(path) >= 0 :
716	>	exe = string.replace(exeWithPath, path,'')
717	>	tar.add(path+exe,exe)
718	>	else :
719	>	tar.add(exeWithPath,os.path.basename(executable))
720	>	pass
721	>	else:
722	>	# the exe is from release, we'll find it on WN
723	>	pass
724	>
725	>	## Now get the libraries: only those in local working area
726	>	libDir = 'lib'
727	>	lib = swArea+'/' +libDir
728	>	common.logger.debug(5,"lib "+lib+" to be tarred")
729	>	if os.path.exists(lib):
730	>	tar.add(lib,libDir)
731	>
732	>	## Now check if module dir is present
733	>	moduleDir = 'module'
734	>	module = swArea + '/' + moduleDir
735	>	if os.path.isdir(module):
736	>	tar.add(module,moduleDir)
737	>
738	>	## Now check if any data dir(s) is present
739	>	swAreaLen=len(swArea)
740	>	self.dataExist = False
741	>	for root, dirs, files in os.walk(swArea):
742	>	if "data" in dirs:
743	>	self.dataExist=True
744	>	common.logger.debug(5,"data "+root+"/data"+" to be tarred")
745	>	tar.add(root+"/data",root[swAreaLen:]+"/data")
746	>
747	>	### CMSSW ParameterSet
748	>	if not self.pset is None:
749	>	cfg_file = common.work_space.jobDir()+self.configFilename()
750	>	tar.add(cfg_file,self.configFilename())
751	>	common.logger.debug(5,"File added to "+self.tgzNameWithPath+" : "+str(tar.getnames()))
752	>
753	>
754	>	## Add ProdCommon dir to tar
755	>	prodcommonDir = 'ProdCommon'
756	>	prodcommonPath = os.environ['CRABDIR'] + '/' + 'external/'
757	>	if os.path.isdir(prodcommonPath):
758	>	tar.add(prodcommonPath,prodcommonDir)
759	>	common.logger.debug(5,"Files added to "+self.tgzNameWithPath+" : "+str(tar.getnames()))
760	>
761	>	##### ML stuff
762	>	ML_file_list=['report.py', 'DashboardAPI.py', 'Logger.py', 'ProcInfo.py', 'apmon.py']
763	>	path=os.environ['CRABDIR'] + '/python/'
764	>	for file in ML_file_list:
765	>	tar.add(path+file,file)
766	>	common.logger.debug(5,"Files added to "+self.tgzNameWithPath+" : "+str(tar.getnames()))
767	>
768	>	##### Utils
769	>	Utils_file_list=['parseCrabFjr.py','writeCfg.py', 'fillCrabFjr.py']
770	>	for file in Utils_file_list:
771	>	tar.add(path+file,file)
772	>	common.logger.debug(5,"Files added to "+self.tgzNameWithPath+" : "+str(tar.getnames()))
773	>
774	>	##### AdditionalFiles
775	>	for file in self.additional_inbox_files:
776	>	tar.add(file,string.split(file,'/')[-1])
777	>	common.logger.debug(5,"Files added to "+self.tgzNameWithPath+" : "+str(tar.getnames()))
778	>
779	>	tar.close()
780	>	except :
781	>	raise CrabException('Could not create tar-ball')
782	>
783	>	## check for tarball size
784	>	tarballinfo = os.stat(self.tgzNameWithPath)
785	>	if ( tarballinfo.st_size > self.MaxTarBallSize*1024*1024 ) :
786	>	raise CrabException('Input sandbox size of ' + str(float(tarballinfo.st_size)/1024.0/1024.0) + ' MB is larger than the allowed ' + str(self.MaxTarBallSize) + ' MB input sandbox limit and not supported by the used GRID submission system. Please make sure that no unnecessary files are in all data directories in your local CMSSW project area as they are automatically packed into the input sandbox.')
787	>
788	>	## create tar-ball with ML stuff
789	>
790	>	def wsSetupEnvironment(self, nj=0):
791		"""
792		Returns part of a job script which prepares
793		the execution environment for the job 'nj'.
794		"""
795	+	if (self.CMSSW_major >= 2 and self.CMSSW_minor >= 1) or (self.CMSSW_major >= 3):
796	+	psetName = 'pset.py'
797	+	else:
798	+	psetName = 'pset.cfg'
799		# Prepare JobType-independent part
800	<	txt = self.wsSetupCMSEnvironment_()
800	>	txt = '\n#Written by cms_cmssw::wsSetupEnvironment\n'
801	>	txt += 'echo ">>> setup environment"\n'
802	>	txt += 'if [ $middleware == LCG ]; then \n'
803	>	txt += self.wsSetupCMSLCGEnvironment_()
804	>	txt += 'elif [ $middleware == OSG ]; then\n'
805	>	txt += '    WORKING_DIR=`/bin/mktemp  -d $OSG_WN_TMP/cms_XXXXXXXXXXXX`\n'
806	>	txt += '    if [ ! $? == 0 ] ;then\n'
807	>	txt += '        echo "ERROR ==> OSG $WORKING_DIR could not be created on WN `hostname`"\n'
808	>	txt += '        job_exit_code=10016\n'
809	>	txt += '        func_exit\n'
810	>	txt += '    fi\n'
811	>	txt += '    echo ">>> Created working directory: $WORKING_DIR"\n'
812	>	txt += '\n'
813	>	txt += '    echo "Change to working directory: $WORKING_DIR"\n'
814	>	txt += '    cd $WORKING_DIR\n'
815	>	txt += '    echo ">>> current directory (WORKING_DIR): $WORKING_DIR"\n'
816	>	txt += self.wsSetupCMSOSGEnvironment_()
817	>	txt += 'fi\n'
818
819		# Prepare JobType-specific part
820		scram = self.scram.commandName()
821		txt += '\n\n'
822	<	txt += 'echo "### SPECIFIC JOB SETUP ENVIRONMENT ###"\n'
822	>	txt += 'echo ">>> specific cmssw setup environment:"\n'
823	>	txt += 'echo "CMSSW_VERSION =  '+self.version+'"\n'
824		txt += scram+' project CMSSW '+self.version+'\n'
825		txt += 'status=$?\n'
826		txt += 'if [ $status != 0 ] ; then\n'
827	<	txt += '   echo "SET_EXE_ENV 1 ==>ERROR CMSSW '+self.version+' not found on `hostname`" \n'
828	<	txt += '   echo "JOB_EXIT_STATUS = 5"\n'
829	<	txt += '   echo "SanityCheckCode = 5" | tee -a $RUNTIME_AREA/$repo\n'
369	<	txt += '   dumpStatus $RUNTIME_AREA/$repo\n'
370	<	txt += '   exit 5 \n'
827	>	txt += '    echo "ERROR ==> CMSSW '+self.version+' not found on `hostname`" \n'
828	>	txt += '    job_exit_code=10034\n'
829	>	txt += '    func_exit\n'
830		txt += 'fi \n'
372	–	txt += 'echo "CMSSW_VERSION =  '+self.version+'"\n'
831		txt += 'cd '+self.version+'\n'
832	<	### needed grep for bug in scramv1 ###
832	>	txt += 'SOFTWARE_DIR=`pwd`\n'
833	>	txt += 'echo ">>> current directory (SOFTWARE_DIR): $SOFTWARE_DIR" \n'
834		txt += 'eval `'+scram+' runtime -sh | grep -v SCRAMRT_LSB_JOBNAME`\n'
835	<
835	>	txt += 'if [ $? != 0 ] ; then\n'
836	>	txt += '    echo "ERROR ==> Problem with the command: "\n'
837	>	txt += '    echo "eval \`'+scram+' runtime -sh | grep -v SCRAMRT_LSB_JOBNAME \` at `hostname`"\n'
838	>	txt += '    job_exit_code=10034\n'
839	>	txt += '    func_exit\n'
840	>	txt += 'fi \n'
841		# Handle the arguments:
842		txt += "\n"
843	<	txt += "## ARGUMNETS: $1 Job Number\n"
380	<	# txt += "## ARGUMNETS: $2 First Event for this job\n"
381	<	# txt += "## ARGUMNETS: $3 Max Event for this job\n"
843	>	txt += "## number of arguments (first argument always jobnumber)\n"
844		txt += "\n"
845	<	txt += "narg=$#\n"
384	<	txt += "if [ $narg -lt 1 ]\n"
845	>	txt += "if [ $nargs -lt "+str(self.argsList)+" ]\n"
846		txt += "then\n"
847	<	txt += "    echo 'SET_EXE_ENV 1 ==> ERROR Too few arguments' +$narg+ \n"
848	<	txt += '    echo "JOB_EXIT_STATUS = 1"\n'
849	<	txt += '    echo "SanityCheckCode = 1" | tee -a $RUNTIME_AREA/$repo\n'
389	<	txt += '    dumpStatus $RUNTIME_AREA/$repo\n'
390	<	txt += "    exit 1\n"
847	>	txt += "    echo 'ERROR ==> Too few arguments' +$nargs+ \n"
848	>	txt += '    job_exit_code=50113\n'
849	>	txt += "    func_exit\n"
850		txt += "fi\n"
851		txt += "\n"
393	–	txt += "NJob=$1\n"
394	–	# txt += "FirstEvent=$2\n"
395	–	# txt += "MaxEvents=$3\n"
852
853		# Prepare job-specific part
854		job = common.job_list[nj]
855	<	pset = os.path.basename(job.configFilename())
856	<	txt += '\n'
857	<	txt += 'cp $RUNTIME_AREA/'+pset+' pset.cfg\n'
858	<	# txt += 'if [ -e $RUNTIME_AREA/orcarc_$CE ] ; then\n'
859	<	# txt += '  cat $RUNTIME_AREA/orcarc_$CE .orcarc >> .orcarc_tmp\n'
860	<	# txt += '  mv .orcarc_tmp .orcarc\n'
861	<	# txt += 'fi\n'
862	<	# txt += 'if [ -e $RUNTIME_AREA/init_$CE.sh ] ; then\n'
863	<	# txt += '  cp $RUNTIME_AREA/init_$CE.sh init.sh\n'
864	<	# txt += 'fi\n'
855	>	if (self.datasetPath):
856	>	txt += '\n'
857	>	txt += 'DatasetPath='+self.datasetPath+'\n'
858	>
859	>	datasetpath_split = self.datasetPath.split("/")
860	>
861	>	txt += 'PrimaryDataset='+datasetpath_split[1]+'\n'
862	>	txt += 'DataTier='+datasetpath_split[2]+'\n'
863	>	txt += 'ApplicationFamily=cmsRun\n'
864	>
865	>	else:
866	>	txt += 'DatasetPath=MCDataTier\n'
867	>	txt += 'PrimaryDataset=null\n'
868	>	txt += 'DataTier=null\n'
869	>	txt += 'ApplicationFamily=MCDataTier\n'
870	>	if self.pset != None:
871	>	pset = os.path.basename(job.configFilename())
872	>	txt += '\n'
873	>	txt += 'cp  $RUNTIME_AREA/'+pset+' .\n'
874	>	if (self.datasetPath): # standard job
875	>	txt += 'InputFiles=${args[1]}; export InputFiles\n'
876	>	if (self.useParent):
877	>	txt += 'ParentFiles=${args[2]}; export ParentFiles\n'
878	>	txt += 'MaxEvents=${args[3]}; export MaxEvents\n'
879	>	txt += 'SkipEvents=${args[4]}; export SkipEvents\n'
880	>	else:
881	>	txt += 'MaxEvents=${args[2]}; export MaxEvents\n'
882	>	txt += 'SkipEvents=${args[3]}; export SkipEvents\n'
883	>	txt += 'echo "Inputfiles:<$InputFiles>"\n'
884	>	if (self.useParent): txt += 'echo "ParentFiles:<$ParentFiles>"\n'
885	>	txt += 'echo "MaxEvents:<$MaxEvents>"\n'
886	>	txt += 'echo "SkipEvents:<$SkipEvents>"\n'
887	>	else:  # pythia like job
888	>	txt += 'PreserveSeeds='  + ','.join(self.preserveSeeds)  + '; export PreserveSeeds\n'
889	>	txt += 'IncrementSeeds=' + ','.join(self.incrementSeeds) + '; export IncrementSeeds\n'
890	>	txt += 'echo "PreserveSeeds: <$PreserveSeeds>"\n'
891	>	txt += 'echo "IncrementSeeds:<$IncrementSeeds>"\n'
892	>	if (self.firstRun):
893	>	txt += 'FirstRun=${args[1]}; export FirstRun\n'
894	>	txt += 'echo "FirstRun: <$FirstRun>"\n'
895	>
896	>	txt += 'mv -f ' + pset + ' ' + psetName + '\n'
897	>
898	>
899	>	if self.pset != None:
900	>	# FUTURE: Can simply for 2_1_x and higher
901	>	txt += '\n'
902	>	if self.debug_wrapper==True:
903	>	txt += 'echo "***** cat ' + psetName + ' *********"\n'
904	>	txt += 'cat ' + psetName + '\n'
905	>	txt += 'echo "****** end ' + psetName + ' ********"\n'
906	>	txt += '\n'
907	>	txt += 'PSETHASH=`edmConfigHash < ' + psetName + '` \n'
908	>	txt += 'echo "PSETHASH = $PSETHASH" \n'
909	>	txt += '\n'
910	>	return txt
911
912	<	if len(self.additional_inbox_files) > 0:
912	>	def wsUntarSoftware(self, nj=0):
913	>	"""
914	>	Put in the script the commands to build an executable
915	>	or a library.
916	>	"""
917	>
918	>	txt = '\n#Written by cms_cmssw::wsUntarSoftware\n'
919	>
920	>	if os.path.isfile(self.tgzNameWithPath):
921	>	txt += 'echo ">>> tar xzvf $RUNTIME_AREA/'+os.path.basename(self.tgzNameWithPath)+' :" \n'
922	>	txt += 'tar xzvf $RUNTIME_AREA/'+os.path.basename(self.tgzNameWithPath)+'\n'
923	>	if  self.debug_wrapper:
924	>	txt += 'ls -Al \n'
925	>	txt += 'untar_status=$? \n'
926	>	txt += 'if [ $untar_status -ne 0 ]; then \n'
927	>	txt += '   echo "ERROR ==> Untarring .tgz file failed"\n'
928	>	txt += '   job_exit_code=$untar_status\n'
929	>	txt += '   func_exit\n'
930	>	txt += 'else \n'
931	>	txt += '   echo "Successful untar" \n'
932	>	txt += 'fi \n'
933	>	txt += '\n'
934	>	txt += 'echo ">>> Include ProdCommon in PYTHONPATH:"\n'
935	>	txt += 'if [ -z "$PYTHONPATH" ]; then\n'
936	>	txt += '   export PYTHONPATH=$RUNTIME_AREA/ProdCommon\n'
937	>	txt += 'else\n'
938	>	txt += '   export PYTHONPATH=$RUNTIME_AREA/ProdCommon:${PYTHONPATH}\n'
939	>	txt += 'echo "PYTHONPATH=$PYTHONPATH"\n'
940	>	txt += 'fi\n'
941	>	txt += '\n'
942	>
943	>	pass
944	>
945	>	return txt
946	>
947	>	def wsBuildExe(self, nj=0):
948	>	"""
949	>	Put in the script the commands to build an executable
950	>	or a library.
951	>	"""
952	>
953	>	txt = '\n#Written by cms_cmssw::wsBuildExe\n'
954	>	txt += 'echo ">>> moving CMSSW software directories in `pwd`" \n'
955	>
956	>	txt += 'rm -r lib/ module/ \n'
957	>	txt += 'mv $RUNTIME_AREA/lib/ . \n'
958	>	txt += 'mv $RUNTIME_AREA/module/ . \n'
959	>	if self.dataExist == True:
960	>	txt += 'rm -r src/ \n'
961	>	txt += 'mv $RUNTIME_AREA/src/ . \n'
962	>	if len(self.additional_inbox_files)>0:
963		for file in self.additional_inbox_files:
964	<	txt += 'if [ -e $RUNTIME_AREA/'+file+' ] ; then\n'
965	<	txt += '   cp $RUNTIME_AREA/'+file+' .\n'
414	<	txt += '   chmod +x '+file+'\n'
415	<	txt += 'fi\n'
416	<	pass
417	<
418	<	# txt += '\n'
419	<	# txt += 'chmod +x ./init.sh\n'
420	<	# txt += './init.sh\n'
421	<	# txt += 'exitStatus=$?\n'
422	<	# txt += 'if [ $exitStatus != 0 ] ; then\n'
423	<	# txt += '  echo "SET_EXE_ENV 1 ==> ERROR StageIn init script failed"\n'
424	<	# txt += '  echo "JOB_EXIT_STATUS = $exitStatus" \n'
425	<	# txt += '  echo "SanityCheckCode = $exitStatus" | tee -a $RUNTIME_AREA/$repo\n'
426	<	# txt += '  dumpStatus $RUNTIME_AREA/$repo\n'
427	<	# txt += '  exit $exitStatus\n'
428	<	# txt += 'fi\n'
429	<	# txt += "echo 'SET_EXE_ENV 0 ==> job setup ok'\n"
430	<	txt += 'echo "### END JOB SETUP ENVIRONMENT ###"\n\n'
431	<
432	<	# txt += 'echo "FirstEvent=$FirstEvent" >> .orcarc\n'
433	<	# txt += 'echo "MaxEvents=$MaxEvents" >> .orcarc\n'
434	<	# if self.ML:
435	<	#     txt += 'echo "MonalisaJobId=$NJob" >> .orcarc\n'
964	>	txt += 'mv $RUNTIME_AREA/'+os.path.basename(file)+' . \n'
965	>	txt += 'mv $RUNTIME_AREA/ProdCommon/ . \n'
966
967	+	txt += 'if [ -z "$PYTHONPATH" ]; then\n'
968	+	txt += '   export PYTHONPATH=$SOFTWARE_DIR/ProdCommon\n'
969	+	txt += 'else\n'
970	+	txt += '   export PYTHONPATH=$SOFTWARE_DIR/ProdCommon:${PYTHONPATH}\n'
971	+	txt += 'echo "PYTHONPATH=$PYTHONPATH"\n'
972	+	txt += 'fi\n'
973		txt += '\n'
974	<	txt += 'echo "***** cat pset.cfg *********"\n'
439	<	txt += 'cat pset.cfg\n'
440	<	txt += 'echo "****** end pset.cfg ********"\n'
974	>
975		return txt
976
977		def modifySteeringCards(self, nj):
978		"""
979	<	modify the card provided by the user,
979	>	modify the card provided by the user,
980		writing a new card into share dir
981		"""
982	<
982	>
983		def executableName(self):
984	<	return self.executable
984	>	if self.scriptExe:
985	>	return "sh "
986	>	else:
987	>	return self.executable
988
989		def executableArgs(self):
990	<	return "-p pset.cfg"
990	>	# FUTURE: This function tests the CMSSW version. Can be simplified as we drop support for old versions
991	>	if self.scriptExe:#CarlosDaniele
992	>	return   self.scriptExe + " $NJob"
993	>	else:
994	>	ex_args = ""
995	>	# FUTURE: This tests the CMSSW version. Can remove code as versions deprecated
996	>	# Framework job report
997	>	if (self.CMSSW_major >= 1 and self.CMSSW_minor >= 5) or (self.CMSSW_major >= 2):
998	>	ex_args += " -j $RUNTIME_AREA/crab_fjr_$NJob.xml"
999	>	# Type of config file
1000	>	if self.CMSSW_major >= 2 :
1001	>	ex_args += " -p pset.py"
1002	>	else:
1003	>	ex_args += " -p pset.cfg"
1004	>	return ex_args
1005
1006		def inputSandbox(self, nj):
1007		"""
1008		Returns a list of filenames to be put in JDL input sandbox.
1009		"""
1010		inp_box = []
460	–	# dict added to delete duplicate from input sandbox file list
461	–	seen = {}
462	–	## code
1011		if os.path.isfile(self.tgzNameWithPath):
1012		inp_box.append(self.tgzNameWithPath)
1013	<	## config
1014	<	inp_box.append(common.job_list[nj].configFilename())
467	<	## additional input files
468	<	for file in self.additional_inbox_files:
469	<	inp_box.append(common.work_space.cwdDir()+file)
470	<	#print "sono inputSandbox, inp_box = ", inp_box
1013	>	wrapper = os.path.basename(str(common._db.queryTask('scriptName')))
1014	>	inp_box.append(common.work_space.pathForTgz() +'job/'+ wrapper)
1015		return inp_box
1016
1017		def outputSandbox(self, nj):
#	Line 476 | Line 1020 | class Cmssw(JobType):
1020		"""
1021		out_box = []
1022
479	–	stdout=common.job_list[nj].stdout()
480	–	stderr=common.job_list[nj].stderr()
481	–
1023		## User Declared output files
1024	<	for out in self.output_file:
1025	<	n_out = nj + 1
1024	>	for out in (self.output_file+self.output_file_sandbox):
1025	>	n_out = nj + 1
1026		out_box.append(self.numberFile_(out,str(n_out)))
1027		return out_box
487	–	return []
1028
1029		def prepareSteeringCards(self):
1030		"""
1031		Make initial modifications of the user's steering card file.
1032		"""
493	–	infile = open(self.pset,'r')
494	–
495	–	outfile = open(common.work_space.jobDir()+self.name()+'.cfg', 'w')
496	–
497	–	outfile.write('\n\n##### The following cards have been created by CRAB: DO NOT TOUCH #####\n')
498	–
499	–	outfile.write('InputCollections=/System/'+self.owner+'/'+self.dataset+'/'+self.dataset+'\n')
500	–
501	–	infile.close()
502	–	outfile.close()
1033		return
1034
1035		def wsRenameOutput(self, nj):
#	Line 507 | Line 1037 | class Cmssw(JobType):
1037		Returns part of a job script which renames the produced files.
1038		"""
1039
1040	<	txt = '\n'
1041	<	file_list = ''
1042	<	for fileWithSuffix in self.output_file:
1040	>	txt = '\n#Written by cms_cmssw::wsRenameOutput\n'
1041	>	txt += 'echo ">>> current directory (SOFTWARE_DIR): $SOFTWARE_DIR" \n'
1042	>	txt += 'echo ">>> current directory content:"\n'
1043	>	if self.debug_wrapper:
1044	>	txt += 'ls -Al\n'
1045	>	txt += '\n'
1046	>
1047	>	for fileWithSuffix in (self.output_file):
1048		output_file_num = self.numberFile_(fileWithSuffix, '$NJob')
514	–	file_list=file_list+output_file_num+' '
515	–	txt += '\n'
516	–	txt += 'ls \n'
1049		txt += '\n'
1050	<	txt += 'ls '+fileWithSuffix+'\n'
1051	<	txt += 'exe_result=$?\n'
1052	<	txt += 'if [ $exe_result -ne 0 ] ; then\n'
1053	<	txt += '   echo "ERROR: No output file to manage"\n'
1054	<	txt += '   echo "JOB_EXIT_STATUS = $exe_result"\n'
1055	<	txt += '   echo "SanityCheckCode = $exe_result" | tee -a $RUNTIME_AREA/$repo\n'
1056	<	txt += '   dumpStatus $RUNTIME_AREA/$repo\n'
1057	<	txt += '   exit $exe_result \n'
1050	>	txt += '# check output file\n'
1051	>	txt += 'if [ -e ./'+fileWithSuffix+' ] ; then\n'
1052	>	if (self.copy_data == 1):  # For OSG nodes, file is in $WORKING_DIR, should not be moved to $RUNTIME_AREA
1053	>	txt += '    mv '+fileWithSuffix+' '+output_file_num+'\n'
1054	>	#txt += '    ln -s `pwd`/'+output_file_num+' $RUNTIME_AREA/'+fileWithSuffix+'\n'
1055	>	else:
1056	>	txt += '    mv '+fileWithSuffix+' $RUNTIME_AREA/'+output_file_num+'\n'
1057	>	txt += '    ln -s $RUNTIME_AREA/'+output_file_num+' $RUNTIME_AREA/'+fileWithSuffix+'\n'
1058		txt += 'else\n'
1059	<	txt += '   cp '+fileWithSuffix+' $RUNTIME_AREA/'+output_file_num+'\n'
1059	>	txt += '    job_exit_code=60302\n'
1060	>	txt += '    echo "WARNING: Output file '+fileWithSuffix+' not found"\n'
1061	>	if common.scheduler.name().upper() == 'CONDOR_G':
1062	>	txt += '    if [ $middleware == OSG ]; then \n'
1063	>	txt += '        echo "prepare dummy output file"\n'
1064	>	txt += '        echo "Processing of job output failed" > $RUNTIME_AREA/'+output_file_num+'\n'
1065	>	txt += '    fi \n'
1066		txt += 'fi\n'
1067	<	txt += 'cd $RUNTIME_AREA\n'
1068	<
1069	<	pass
1070	<
1071	<	file_list=file_list[:-1]
1072	<	txt += 'file_list="'+file_list+'"\n'
1067	>	file_list = []
1068	>	for fileWithSuffix in (self.output_file):
1069	>	file_list.append(self.numberFile_(fileWithSuffix, '$NJob'))
1070	>
1071	>	txt += 'file_list="'+string.join(file_list,' ')+'"\n'
1072	>	txt += '\n'
1073	>	txt += 'echo ">>> current directory (SOFTWARE_DIR): $SOFTWARE_DIR" \n'
1074	>	txt += 'echo ">>> current directory content:"\n'
1075	>	if self.debug_wrapper:
1076	>	txt += 'ls -Al\n'
1077	>	txt += '\n'
1078	>	txt += 'cd $RUNTIME_AREA\n'
1079	>	txt += 'echo ">>> current directory (RUNTIME_AREA):  $RUNTIME_AREA"\n'
1080		return txt
1081
1082		def numberFile_(self, file, txt):
#	Line 542 | Line 1087 | class Cmssw(JobType):
1087		# take away last extension
1088		name = p[0]
1089		for x in p[1:-1]:
1090	<	name=name+"."+x
1090	>	name=name+"."+x
1091		# add "_txt"
1092		if len(p)>1:
1093	<	ext = p[len(p)-1]
1094	<	#result = name + '_' + str(txt) + "." + ext
1095	<	result = name + '_' + txt + "." + ext
1096	<	else:
1097	<	#result = name + '_' + str(txt)
553	<	result = name + '_' + txt
554	<
1093	>	ext = p[len(p)-1]
1094	>	result = name + '_' + txt + "." + ext
1095	>	else:
1096	>	result = name + '_' + txt
1097	>
1098		return result
1099
1100	<	def getRequirements(self):
1100	>	def getRequirements(self, nj=[]):
1101		"""
1102	<	return job requirements to add to jdl files
1102	>	return job requirements to add to jdl files
1103		"""
1104		req = ''
1105	<	if common.analisys_common_info['sites']:
1106	<	if common.analisys_common_info['sw_version']:
1107	<	req='Member("VO-cms-' + \
1108	<	common.analisys_common_info['sw_version'] + \
1109	<	'", other.GlueHostApplicationSoftwareRunTimeEnvironment)'
1110	<	if len(common.analisys_common_info['sites'])>0:
1111	<	req = req + ' && ('
1112	<	for i in range(len(common.analisys_common_info['sites'])):
1113	<	req = req + 'other.GlueCEInfoHostName == "' \
1114	<	+ common.analisys_common_info['sites'][i] + '"'
1115	<	if ( i < (int(len(common.analisys_common_info['sites']) - 1)) ):
1116	<	req = req + ' || '
1117	<	req = req + ')'
575	<	#print "req = ", req
1105	>	if self.version:
1106	>	req='Member("VO-cms-' + \
1107	>	self.version + \
1108	>	'", other.GlueHostApplicationSoftwareRunTimeEnvironment)'
1109	>	if self.executable_arch:
1110	>	req+=' && Member("VO-cms-' + \
1111	>	self.executable_arch + \
1112	>	'", other.GlueHostApplicationSoftwareRunTimeEnvironment)'
1113	>
1114	>	req = req + ' && (other.GlueHostNetworkAdapterOutboundIP)'
1115	>	if common.scheduler.name() == "glitecoll":
1116	>	req += ' && other.GlueCEStateStatus == "Production" '
1117	>
1118		return req
1119	+
1120	+	def configFilename(self):
1121	+	""" return the config filename """
1122	+	# FUTURE: Can remove cfg mode for CMSSW >= 2_1_x
1123	+	if (self.CMSSW_major >= 2 and self.CMSSW_minor >= 1) or (self.CMSSW_major >= 3):
1124	+	return self.name()+'.py'
1125	+	else:
1126	+	return self.name()+'.cfg'
1127	+
1128	+	def wsSetupCMSOSGEnvironment_(self):
1129	+	"""
1130	+	Returns part of a job script which is prepares
1131	+	the execution environment and which is common for all CMS jobs.
1132	+	"""
1133	+	txt = '\n#Written by cms_cmssw::wsSetupCMSOSGEnvironment_\n'
1134	+	txt += '    echo ">>> setup CMS OSG environment:"\n'
1135	+	txt += '    echo "set SCRAM ARCH to ' + self.executable_arch + '"\n'
1136	+	txt += '    export SCRAM_ARCH='+self.executable_arch+'\n'
1137	+	txt += '    echo "SCRAM_ARCH = $SCRAM_ARCH"\n'
1138	+	txt += '    if [ -f $OSG_APP/cmssoft/cms/cmsset_default.sh ] ;then\n'
1139	+	txt += '      # Use $OSG_APP/cmssoft/cms/cmsset_default.sh to setup cms software\n'
1140	+	txt += '        source $OSG_APP/cmssoft/cms/cmsset_default.sh '+self.version+'\n'
1141	+	txt += '    else\n'
1142	+	txt += '        echo "ERROR ==> $OSG_APP/cmssoft/cms/cmsset_default.sh file not found"\n'
1143	+	txt += '        job_exit_code=10020\n'
1144	+	txt += '        func_exit\n'
1145	+	txt += '    fi\n'
1146	+	txt += '\n'
1147	+	txt += '    echo "==> setup cms environment ok"\n'
1148	+	txt += '    echo "SCRAM_ARCH = $SCRAM_ARCH"\n'
1149	+
1150	+	return txt
1151	+
1152	+	def wsSetupCMSLCGEnvironment_(self):
1153	+	"""
1154	+	Returns part of a job script which is prepares
1155	+	the execution environment and which is common for all CMS jobs.
1156	+	"""
1157	+	txt = '\n#Written by cms_cmssw::wsSetupCMSLCGEnvironment_\n'
1158	+	txt += '    echo ">>> setup CMS LCG environment:"\n'
1159	+	txt += '    echo "set SCRAM ARCH and BUILD_ARCH to ' + self.executable_arch + ' ###"\n'
1160	+	txt += '    export SCRAM_ARCH='+self.executable_arch+'\n'
1161	+	txt += '    export BUILD_ARCH='+self.executable_arch+'\n'
1162	+	txt += '    if [ ! $VO_CMS_SW_DIR ] ;then\n'
1163	+	txt += '        echo "ERROR ==> CMS software dir not found on WN `hostname`"\n'
1164	+	txt += '        job_exit_code=10031\n'
1165	+	txt += '        func_exit\n'
1166	+	txt += '    else\n'
1167	+	txt += '        echo "Sourcing environment... "\n'
1168	+	txt += '        if [ ! -s $VO_CMS_SW_DIR/cmsset_default.sh ] ;then\n'
1169	+	txt += '            echo "ERROR ==> cmsset_default.sh file not found into dir $VO_CMS_SW_DIR"\n'
1170	+	txt += '            job_exit_code=10020\n'
1171	+	txt += '            func_exit\n'
1172	+	txt += '        fi\n'
1173	+	txt += '        echo "sourcing $VO_CMS_SW_DIR/cmsset_default.sh"\n'
1174	+	txt += '        source $VO_CMS_SW_DIR/cmsset_default.sh\n'
1175	+	txt += '        result=$?\n'
1176	+	txt += '        if [ $result -ne 0 ]; then\n'
1177	+	txt += '            echo "ERROR ==> problem sourcing $VO_CMS_SW_DIR/cmsset_default.sh"\n'
1178	+	txt += '            job_exit_code=10032\n'
1179	+	txt += '            func_exit\n'
1180	+	txt += '        fi\n'
1181	+	txt += '    fi\n'
1182	+	txt += '    \n'
1183	+	txt += '    echo "==> setup cms environment ok"\n'
1184	+	return txt
1185	+
1186	+	def modifyReport(self, nj):
1187	+	"""
1188	+	insert the part of the script that modifies the FrameworkJob Report
1189	+	"""
1190	+	txt = '\n#Written by cms_cmssw::modifyReport\n'
1191	+	publish_data = int(self.cfg_params.get('USER.publish_data',0))
1192	+	if (publish_data == 1):
1193	+	processedDataset = self.cfg_params['USER.publish_data_name']
1194	+	LFNBaseName = LFNBase(processedDataset)
1195	+
1196	+	txt += 'if [ $copy_exit_status -eq 0 ]; then\n'
1197	+	txt += '    FOR_LFN=%s_${PSETHASH}/\n'%(LFNBaseName)
1198	+	txt += 'else\n'
1199	+	txt += '    FOR_LFN=/copy_problems/ \n'
1200	+	txt += '    SE=""\n'
1201	+	txt += '    SE_PATH=""\n'
1202	+	txt += 'fi\n'
1203	+
1204	+	txt += 'echo ">>> Modify Job Report:" \n'
1205	+	txt += 'chmod a+x $SOFTWARE_DIR/ProdCommon/ProdCommon/FwkJobRep/ModifyJobReport.py\n'
1206	+	txt += 'ProcessedDataset='+processedDataset+'\n'
1207	+	txt += 'echo "ProcessedDataset = $ProcessedDataset"\n'
1208	+	txt += 'echo "SE = $SE"\n'
1209	+	txt += 'echo "SE_PATH = $SE_PATH"\n'
1210	+	txt += 'echo "FOR_LFN = $FOR_LFN" \n'
1211	+	txt += 'echo "CMSSW_VERSION = $CMSSW_VERSION"\n\n'
1212	+	txt += 'echo "$SOFTWARE_DIR/ProdCommon/ProdCommon/FwkJobRep/ModifyJobReport.py $RUNTIME_AREA/crab_fjr_$NJob.xml $NJob $FOR_LFN $PrimaryDataset $DataTier $ProcessedDataset $ApplicationFamily $executable $CMSSW_VERSION $PSETHASH $SE $SE_PATH"\n'
1213	+	txt += '$SOFTWARE_DIR/ProdCommon/ProdCommon/FwkJobRep/ModifyJobReport.py $RUNTIME_AREA/crab_fjr_$NJob.xml $NJob $FOR_LFN $PrimaryDataset $DataTier $ProcessedDataset $ApplicationFamily $executable $CMSSW_VERSION $PSETHASH $SE $SE_PATH\n'
1214	+	txt += 'modifyReport_result=$?\n'
1215	+	txt += 'if [ $modifyReport_result -ne 0 ]; then\n'
1216	+	txt += '    modifyReport_result=70500\n'
1217	+	txt += '    job_exit_code=$modifyReport_result\n'
1218	+	txt += '    echo "ModifyReportResult=$modifyReport_result" | tee -a $RUNTIME_AREA/$repo\n'
1219	+	txt += '    echo "WARNING: Problem with ModifyJobReport"\n'
1220	+	txt += 'else\n'
1221	+	txt += '    mv NewFrameworkJobReport.xml $RUNTIME_AREA/crab_fjr_$NJob.xml\n'
1222	+	txt += 'fi\n'
1223	+	return txt
1224	+
1225	+	def wsParseFJR(self):
1226	+	"""
1227	+	Parse the FrameworkJobReport to obtain useful infos
1228	+	"""
1229	+	txt = '\n#Written by cms_cmssw::wsParseFJR\n'
1230	+	txt += 'echo ">>> Parse FrameworkJobReport crab_fjr.xml"\n'
1231	+	txt += 'if [ -s $RUNTIME_AREA/crab_fjr_$NJob.xml ]; then\n'
1232	+	txt += '    if [ -s $RUNTIME_AREA/parseCrabFjr.py ]; then\n'
1233	+	txt += '        cmd_out=`python $RUNTIME_AREA/parseCrabFjr.py --input $RUNTIME_AREA/crab_fjr_$NJob.xml --dashboard $MonitorID,$MonitorJobID '+self.debugWrap+'`\n'
1234	+	if self.debug_wrapper :
1235	+	txt += '        echo "Result of parsing the FrameworkJobReport crab_fjr.xml: $cmd_out"\n'
1236	+	txt += '        executable_exit_status=`python $RUNTIME_AREA/parseCrabFjr.py --input $RUNTIME_AREA/crab_fjr_$NJob.xml --exitcode`\n'
1237	+	txt += '        if [ $executable_exit_status -eq 50115 ];then\n'
1238	+	txt += '            echo ">>> crab_fjr.xml contents: "\n'
1239	+	txt += '            cat $RUNTIME_AREA/crab_fjr_NJob.xml\n'
1240	+	txt += '            echo "Wrong FrameworkJobReport --> does not contain useful info. ExitStatus: $executable_exit_status"\n'
1241	+	txt += '        elif [ $executable_exit_status -eq -999 ];then\n'
1242	+	txt += '            echo "ExitStatus from FrameworkJobReport not available. not available. Using exit code of executable from command line."\n'
1243	+	txt += '        else\n'
1244	+	txt += '            echo "Extracted ExitStatus from FrameworkJobReport parsing output: $executable_exit_status"\n'
1245	+	txt += '        fi\n'
1246	+	txt += '    else\n'
1247	+	txt += '        echo "CRAB python script to parse CRAB FrameworkJobReport crab_fjr.xml is not available, using exit code of executable from command line."\n'
1248	+	txt += '    fi\n'
1249	+	#### Patch to check input data reading for CMSSW16x Hopefully we-ll remove it asap
1250	+
1251	+	if self.datasetPath:
1252	+	# VERIFY PROCESSED DATA
1253	+	txt += '    if [ $executable_exit_status -eq 0 ];then\n'
1254	+	txt += '      echo ">>> Verify list of processed files:"\n'
1255	+	txt += '      echo $InputFiles |tr -d \'\\\\\' |tr \',\' \'\\n\'|tr -d \'"\' > input-files.txt\n'
1256	+	txt += '      python $RUNTIME_AREA/parseCrabFjr.py --input $RUNTIME_AREA/crab_fjr_$NJob.xml --lfn > processed-files.txt\n'
1257	+	txt += '      cat input-files.txt  | sort | uniq > tmp.txt\n'
1258	+	txt += '      mv tmp.txt input-files.txt\n'
1259	+	txt += '      echo "cat input-files.txt"\n'
1260	+	txt += '      echo "----------------------"\n'
1261	+	txt += '      cat input-files.txt\n'
1262	+	txt += '      cat processed-files.txt | sort | uniq > tmp.txt\n'
1263	+	txt += '      mv tmp.txt processed-files.txt\n'
1264	+	txt += '      echo "----------------------"\n'
1265	+	txt += '      echo "cat processed-files.txt"\n'
1266	+	txt += '      echo "----------------------"\n'
1267	+	txt += '      cat processed-files.txt\n'
1268	+	txt += '      echo "----------------------"\n'
1269	+	txt += '      diff -q input-files.txt processed-files.txt\n'
1270	+	txt += '      fileverify_status=$?\n'
1271	+	txt += '      if [ $fileverify_status -ne 0 ]; then\n'
1272	+	txt += '         executable_exit_status=30001\n'
1273	+	txt += '         echo "ERROR ==> not all input files processed"\n'
1274	+	txt += '         echo "      ==> list of processed files from crab_fjr.xml differs from list in pset.cfg"\n'
1275	+	txt += '         echo "      ==> diff input-files.txt processed-files.txt"\n'
1276	+	txt += '      fi\n'
1277	+	txt += '    fi\n'
1278	+	txt += '\n'
1279	+	txt += 'else\n'
1280	+	txt += '    echo "CRAB FrameworkJobReport crab_fjr.xml is not available, using exit code of executable from command line."\n'
1281	+	txt += 'fi\n'
1282	+	txt += '\n'
1283	+	txt += 'echo "ExeExitCode=$executable_exit_status" | tee -a $RUNTIME_AREA/$repo\n'
1284	+	txt += 'echo "EXECUTABLE_EXIT_STATUS = $executable_exit_status"\n'
1285	+	txt += 'job_exit_code=$executable_exit_status\n'
1286	+
1287	+	return txt
1288	+
1289	+	def setParam_(self, param, value):
1290	+	self._params[param] = value
1291	+
1292	+	def getParams(self):
1293	+	return self._params
1294	+
1295	+	def uniquelist(self, old):
1296	+	"""
1297	+	remove duplicates from a list
1298	+	"""
1299	+	nd={}
1300	+	for e in old:
1301	+	nd[e]=0
1302	+	return nd.keys()
1303	+
1304	+	def outList(self):
1305	+	"""
1306	+	check the dimension of the output files
1307	+	"""
1308	+	txt = ''
1309	+	txt += 'echo ">>> list of expected files on output sandbox"\n'
1310	+	listOutFiles = []
1311	+	stdout = 'CMSSW_$NJob.stdout'
1312	+	stderr = 'CMSSW_$NJob.stderr'
1313	+	if (self.return_data == 1):
1314	+	for file in (self.output_file+self.output_file_sandbox):
1315	+	listOutFiles.append(self.numberFile_(file, '$NJob'))
1316	+	listOutFiles.append(stdout)
1317	+	listOutFiles.append(stderr)
1318	+	else:
1319	+	for file in (self.output_file_sandbox):
1320	+	listOutFiles.append(self.numberFile_(file, '$NJob'))
1321	+	listOutFiles.append(stdout)
1322	+	listOutFiles.append(stderr)
1323	+	txt += 'echo "output files: '+string.join(listOutFiles,' ')+'"\n'
1324	+	txt += 'filesToCheck="'+string.join(listOutFiles,' ')+'"\n'
1325	+	txt += 'export filesToCheck\n'
1326	+	return txt

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines