def __call__(self):
valid = False
configmsg = 'Default'
self.logger.debug("Started submission")
serverFactory = CRABClient.Emulator.getEmulator('rest')
uniquerequestname = None
self.logger.debug("Working on %s" % str(self.requestarea))
configreq = {'dryrun': 1 if self.options.dryrun else 0}
for param in parametersMapping['on-server']:
mustbetype = getattr(types, parametersMapping['on-server'][param]['type'])
default = parametersMapping['on-server'][param]['default']
config_params = parametersMapping['on-server'][param]['config']
for config_param in config_params:
attrs = config_param.split('.')
temp = self.configuration
for attr in attrs:
temp = getattr(temp, attr, None)
if temp is None:
break
if temp is not None:
configreq[param] = temp
break
elif default is not None:
configreq[param] = default
temp = default
else:
## Parameter not strictly required.
pass
## Check that the requestname is of the right type.
## This is not checked in SubCommand.validateConfig().
if param == 'workflow':
if mustbetype == type(self.requestname):
configreq['workflow'] = self.requestname
## Translate boolean flags into integers.
elif param in ['savelogsflag', 'publication', 'nonprodsw', 'useparent', 'ignorelocality', 'saveoutput', 'oneEventMode']:
configreq[param] = 1 if temp else 0
## Translate DBS URL aliases into DBS URLs.
elif param in ['dbsurl', 'publishdbsurl']:
if param == 'dbsurl':
dbstype = 'reader'
elif param == 'publishdbsurl':
dbstype = 'writer'
allowed_dbsurls = DBSURLS[dbstype].values()
allowed_dbsurls_aliases = DBSURLS[dbstype].keys()
if configreq[param] in allowed_dbsurls_aliases:
configreq[param] = DBSURLS[dbstype][configreq[param]]
elif configreq[param].rstrip('/') in allowed_dbsurls:
configreq[param] = configreq[param].rstrip('/')
elif param == 'scriptexe' and 'scriptexe' in configreq:
configreq[param] = os.path.basename(configreq[param])
jobconfig = {}
#get the backend URLs from the server external configuration
serverBackendURLs = server_info('backendurls', self.serverurl, self.proxyfilename, getUrl(self.instance, resource='info'))
#if cacheSSL is specified in the server external configuration we will use it to upload the sandbox (baseURL will be ignored)
filecacheurl = serverBackendURLs['cacheSSL'] if 'cacheSSL' in serverBackendURLs else None
pluginParams = [ self.configuration, self.logger, os.path.join(self.requestarea, 'inputs') ]
crab_job_types = getJobTypes()
if upper(configreq['jobtype']) in crab_job_types:
plugjobtype = crab_job_types[upper(configreq['jobtype'])](*pluginParams)
inputfiles, jobconfig, isbchecksum = plugjobtype.run(filecacheurl)
else:
fullname = configreq['jobtype']
basename = os.path.basename(fullname).split('.')[0]
plugin = addPlugin(fullname)[basename]
pluginInst = plugin(*pluginParams)
inputfiles, jobconfig, isbchecksum = pluginInst.run()
if configreq['publication']:
non_edm_files = jobconfig['tfileoutfiles'] + jobconfig['addoutputfiles']
if non_edm_files:
msg = "%sWarning%s: The following output files will not be published, as they are not EDM files: %s" % (colors.RED, colors.NORMAL, non_edm_files)
self.logger.warning(msg)
if not configreq['publishname']:
configreq['publishname'] = isbchecksum
else:
configreq['publishname'] = "%s-%s" %(configreq['publishname'], isbchecksum)
configreq.update(jobconfig)
server = serverFactory(self.serverurl, self.proxyfilename, self.proxyfilename, version=__version__)
self.logger.info("Sending the request to the server")
self.logger.debug("Submitting %s " % str(configreq))
## TODO: this shouldn't be hard-coded.
listParams = ['adduserfiles', 'addoutputfiles', 'sitewhitelist', 'siteblacklist', 'blockwhitelist', 'blockblacklist', \
'tfileoutfiles', 'edmoutfiles', 'runs', 'lumis', 'userfiles', 'scriptargs', 'extrajdl']
configreq_encoded = self._encodeRequest(configreq, listParams)
self.logger.debug('Encoded submit request: %s' % (configreq_encoded))
dictresult, status, reason = server.put( self.uri, data = configreq_encoded)
self.logger.debug("Result: %s" % dictresult)
if status != 200:
msg = "Problem sending the request:\ninput:%s\noutput:%s\nreason:%s" % (str(configreq), str(dictresult), str(reason))
raise RESTCommunicationException(msg)
elif dictresult.has_key("result"):
uniquerequestname = dictresult["result"][0]["RequestName"]
else:
msg = "Problem during submission, no request ID returned:\ninput:%s\noutput:%s\nreason:%s" \
% (str(configreq), str(dictresult), str(reason))
raise RESTCommunicationException(msg)
tmpsplit = self.serverurl.split(':')
createCache(self.requestarea, tmpsplit[0], tmpsplit[1] if len(tmpsplit)>1 else '', uniquerequestname,
voRole=self.voRole, voGroup=self.voGroup, instance=self.instance,
originalConfig = self.configuration)
self.logger.info("%sSuccess%s: Your task has been delivered to the CRAB3 server." %(colors.GREEN, colors.NORMAL))
if not (self.options.wait or self.options.dryrun):
self.logger.info("Task name: %s" % uniquerequestname)
self.logger.info("Please use 'crab status' to check how the submission process proceeds.")
else:
targetTaskStatus = 'UPLOADED' if self.options.dryrun else 'SUBMITTED'
self.checkStatusLoop(server, uniquerequestname, targetTaskStatus)
if self.options.dryrun:
self.printDryRunResults(*self.executeTestRun(filecacheurl))
self.logger.debug("About to return")
return {'requestname' : self.requestname , 'uniquerequestname' : uniquerequestname }
def validateConfig(self):
"""
__validateConfig__
Checking if needed input parameters are there
"""
valid, msg = SubCommand.validateConfig(self)
if not valid:
return False, msg
## Check that Data.unitsPerjob is specified.
if hasattr(self.configuration.Data, 'unitsPerJob'):
try:
float(self.configuration.Data.unitsPerJob)
except ValueError:
msg = "Invalid CRAB configuration: Parameter Data.unitsPerJob must be a valid number, not %s." % (self.configuration.Data.unitsPerJob)
return False, msg
## Check that JobType.pluginName and JobType.externalPluginFile are not both specified.
if hasattr(self.configuration.JobType, 'pluginName') and hasattr(self.configuration.JobType, 'externalPluginFile'):
msg = "Invalid CRAB configuration: Only one of JobType.pluginName or JobType.externalPluginFile parameters can be specified."
pluginName_default = getParamDefaultValue('JobType.pluginName')
if pluginName_default:
msg += "\nIf neither JobType.pluginName nor JobType.externalPluginFile would be specified,"
msg += " the default JobType.pluginName = '%s' would be used." % (pluginName_default)
return False, msg
## Load the external plugin or check that the crab plugin is valid.
external_plugin_name = getattr(self.configuration.JobType, 'externalPluginFile', None)
crab_plugin_name = getattr(self.configuration.JobType, 'pluginName', None)
crab_job_types = {'ANALYSIS': None, 'PRIVATEMC': None} #getJobTypes()
if external_plugin_name:
addPlugin(external_plugin_name) # Do we need to do this here?
if crab_plugin_name:
if upper(crab_plugin_name) not in crab_job_types:
msg = "Invalid CRAB configuration: Parameter JobType.pluginName has an invalid value ('%s')." % (crab_plugin_name)
msg += "\nAllowed values are: %s." % (", ".join(['%s' % job_type for job_type in crab_job_types.keys()]))
return False, msg
msg = "Will use CRAB %s plugin" % ("Analysis" if upper(crab_plugin_name) == 'ANALYSIS' else "PrivateMC")
msg += " (i.e. will run %s job type)." % ("an analysis" if upper(crab_plugin_name) == 'ANALYSIS' else "a MC generation")
self.logger.debug(msg)
## Check that the particular combination (Data.publication = True, General.transferOutputs = False) is not specified.
if getattr(self.configuration.Data, 'publication', getParamDefaultValue('Data.publication')) and \
not getattr(self.configuration.General, 'transferOutputs', getParamDefaultValue('General.transferOutputs')):
msg = "Invalid CRAB configuration: Data.publication is True, but General.transferOutputs is False."
msg += "\nPublication can not be performed if the output files are not transferred to a permanent storage."
return False, msg
## Check that a storage site is specified if General.transferOutputs = True or General.transferLogs = True.
if not hasattr(self.configuration.Site, 'storageSite'):
if getattr(self.configuration.General, 'transferLogs', getParamDefaultValue('General.transferLogs')) or \
getattr(self.configuration.General, 'transferOutputs', getParamDefaultValue('General.transferOutputs')):
msg = "Invalid CRAB configuration: Parameter Site.storageSite is missing."
return False, msg
## If an input dataset and a DBS URL are specified, check that the DBS URL is a good one.
## Also, if the DBS URL is 'phys0x', check that the input dataset tier is USER.
if hasattr(self.configuration.Data, 'inputDBS'):
if hasattr(self.configuration.Data, 'inputDataset'):
msg = None
dbs_urls_aliases = DBSURLS['reader'].keys()
dbs_urls = DBSURLS['reader'].values()
if (self.configuration.Data.inputDBS not in dbs_urls_aliases) and (self.configuration.Data.inputDBS.rstrip('/') not in dbs_urls):
msg = "Invalid CRAB configuration: Parameter Data.inputDBS has an invalid value ('%s')." % (self.configuration.Data.inputDBS)
msg += "\nAllowed values are: "
msg += "\n ".join(["'%s' ('%s')" % (alias, url) for alias, url in DBSURLS['reader'].iteritems()])
local_dbs_urls_aliases = ['phys01', 'phys02', 'phys03']
local_dbs_urls = [DBSURLS['reader'][alias] for alias in local_dbs_urls_aliases if alias in DBSURLS['reader']]
if self.configuration.Data.inputDBS in local_dbs_urls + local_dbs_urls_aliases:
inputDataset_parts = self.configuration.Data.inputDataset.split('/')
inputDataset_parts.pop(0)
inputDataset_tier = inputDataset_parts[-1] if len(inputDataset_parts) == 3 else None
user_data_tiers = ['USER']
if inputDataset_tier not in user_data_tiers:
msg = "Invalid CRAB configuration: A local DBS instance '%s' was specified for reading an input dataset of tier %s." \
% (self.configuration.Data.inputDBS, inputDataset_tier)
msg += "\nDatasets of tier different than %s must be read from the global DBS instance; this is, set Data.inputDBS = 'global'." \
% (", ".join(user_data_tiers[:-1]) + " or " + user_data_tiers[-1] if len(user_data_tiers) > 1 else user_data_tiers[0])
if msg:
inputDBS_default = getParamDefaultValue('Data.inputDBS')
if inputDBS_default:
inputDBS_default, inputDBS_default_alias = self.getDBSURLAndAlias(inputDBS_default, 'reader')
if inputDBS_default and inputDBS_default_alias:
msg += "\nIf Data.inputDBS would not be specified, the default '%s' ('%s') would be used." % (inputDBS_default_alias, inputDBS_default)
return False, msg
## If a publication DBS URL is specified and publication is ON, check that the DBS URL is a good one.
if hasattr(self.configuration.Data, 'publishDBS'):
if getattr(self.configuration.Data, 'publication', getParamDefaultValue('Data.publication')):
dbs_urls = DBSURLS['writer'].values()
dbs_urls_aliases = DBSURLS['writer'].keys()
if (self.configuration.Data.publishDBS not in dbs_urls_aliases) and (self.configuration.Data.publishDBS.rstrip('/') not in dbs_urls):
msg = "Invalid CRAB configuration: Parameter Data.publishDBS has an invalid value ('%s')." % (self.configuration.Data.publishDBS)
msg += "\nAllowed values are: "
msg += "\n ".join(["'%s' ('%s')" % (alias, url) for alias, url in DBSURLS['writer'].iteritems()])
publishDBS_default = getParamDefaultValue('Data.publishDBS')
if publishDBS_default:
publishDBS_default, publishDBS_default_alias = self.getDBSURLAndAlias(publishDBS_default, 'writer')
if publishDBS_default and publishDBS_default_alias:
msg += "\nIf Data.publishDBS would not be specified, the default '%s' ('%s') would be used." \
% (publishDBS_default_alias, publishDBS_default)
return False, msg
if hasattr(self.configuration.JobType, 'scriptExe'):
if not os.path.isfile(self.configuration.JobType.scriptExe):
msg = "Cannot find the file %s specified in the JobType.scriptExe configuration parameter." % (self.configuration.JobType.scriptExe)
return False, msg
return True, "Valid configuration"
def validateConfig(self):
"""
__validateConfig__
Checking if needed input parameters are there
"""
valid, msg = SubCommand.validateConfig(self)
if not valid:
return False, msg
requestNameLenLimit = 100
if hasattr(self.configuration.General, 'requestName'):
if len(self.configuration.General.requestName
) > requestNameLenLimit:
msg = "Invalid CRAB configuration: Parameter General.requestName should not be longer than %d characters." % (
requestNameLenLimit)
return False, msg
splitting = getattr(self.configuration.Data, 'splitting', 'Automatic')
autoSplitt = True if splitting == 'Automatic' else False
autoSplittUnitsMin = 180 # 3 hours (defined also in TW config as 'minAutomaticRuntimeMins')
autoSplittUnitsMax = 2700 # 45 hours
## Check that maxJobRuntimeMin is not used with Automatic splitting
if autoSplitt and hasattr(self.configuration.JobType,
'maxJobRuntimeMin'):
msg = "The 'maxJobRuntimeMin' parameter is not compatible with the 'Automatic' splitting mode (default)."
return False, msg
## Check that --dryrun is not used with Automatic splitting
if autoSplitt and self.options.dryrun:
msg = "The 'dryrun' option is not compatible with the 'Automatic' splitting mode (default)."
return False, msg
## Check that Data.unitsPerjob is specified.
if hasattr(self.configuration.Data, 'unitsPerJob'):
try:
float(self.configuration.Data.unitsPerJob)
except ValueError:
msg = "Invalid CRAB configuration: Parameter Data.unitsPerJob must be a valid number, not %s." % (
self.configuration.Data.unitsPerJob)
return False, msg
if not int(self.configuration.Data.unitsPerJob) > 0:
msg = "Invalid CRAB configuration: Parameter Data.unitsPerJob must be > 0, not %s." % (
self.configuration.Data.unitsPerJob)
return False, msg
if autoSplitt and (
self.configuration.Data.unitsPerJob > autoSplittUnitsMax or
self.configuration.Data.unitsPerJob < autoSplittUnitsMin):
msg = "Invalid CRAB configuration: In case of Automatic splitting, the Data.unitsPerJob parameter must be in the [%d, %d] minutes range. You asked for %d minutes." % (
autoSplittUnitsMin, autoSplittUnitsMax,
self.configuration.Data.unitsPerJob)
return False, msg
elif not autoSplitt:
# The default value is only valid for automatic splitting!
msg = "Invalid CRAB configuration: Parameter Data.unitsPerJob is mandatory for '%s' splitting mode." % splitting
return False, msg
## Check that JobType.pluginName and JobType.externalPluginFile are not both specified.
if hasattr(self.configuration.JobType, 'pluginName') and hasattr(
self.configuration.JobType, 'externalPluginFile'):
msg = "Invalid CRAB configuration: Only one of JobType.pluginName or JobType.externalPluginFile parameters can be specified."
pluginName_default = getParamDefaultValue('JobType.pluginName')
if pluginName_default:
msg += "\nIf neither JobType.pluginName nor JobType.externalPluginFile would be specified,"
msg += " the default JobType.pluginName = '%s' would be used." % (
pluginName_default)
return False, msg
## Load the external plugin or check that the crab plugin is valid.
external_plugin_name = getattr(self.configuration.JobType,
'externalPluginFile', None)
crab_plugin_name = getattr(self.configuration.JobType, 'pluginName',
None)
crab_job_types = {
'ANALYSIS': None,
'PRIVATEMC': None,
'COPYCAT': None
} #getJobTypes()
if external_plugin_name:
addPlugin(external_plugin_name) # Do we need to do this here?
if crab_plugin_name:
if upper(crab_plugin_name) not in crab_job_types:
msg = "Invalid CRAB configuration: Parameter JobType.pluginName has an invalid value ('%s')." % (
crab_plugin_name)
msg += "\nAllowed values are: %s." % (", ".join(
['%s' % job_type for job_type in crab_job_types.keys()]))
return False, msg
msg = "Will use CRAB %s plugin" % ("Analysis" if upper(
crab_plugin_name) == 'ANALYSIS' else "PrivateMC")
msg += " (i.e. will run %s job type)." % ("an analysis" if upper(
crab_plugin_name) == 'ANALYSIS' else "a MC generation")
self.logger.debug(msg)
## Check that the particular combination (Data.publication = True, General.transferOutputs = False) is not specified.
if getattr(self.configuration.Data, 'publication', getParamDefaultValue('Data.publication')) and \
not getattr(self.configuration.General, 'transferOutputs', getParamDefaultValue('General.transferOutputs')):
msg = "Invalid CRAB configuration: Data.publication is True, but General.transferOutputs is False."
msg += "\nPublication can not be performed if the output files are not transferred to a permanent storage."
return False, msg
## Check that a storage site is specified if General.transferOutputs = True or General.transferLogs = True.
if not hasattr(self.configuration.Site, 'storageSite'):
if getattr(self.configuration.General, 'transferLogs', getParamDefaultValue('General.transferLogs')) or \
getattr(self.configuration.General, 'transferOutputs', getParamDefaultValue('General.transferOutputs')):
msg = "Invalid CRAB configuration: Parameter Site.storageSite is missing."
return False, msg
## If an input dataset and a DBS URL are specified, check that the DBS URL is a good one.
## Also, if the DBS URL is 'phys0x', check that the input dataset tier is USER.
if hasattr(self.configuration.Data, 'inputDBS'):
if hasattr(self.configuration.Data, 'inputDataset'):
msg = None
dbs_urls_aliases = DBSURLS['reader'].keys()
dbs_urls = DBSURLS['reader'].values()
if (self.configuration.Data.inputDBS not in dbs_urls_aliases
) and (self.configuration.Data.inputDBS.rstrip('/')
not in dbs_urls):
msg = "Invalid CRAB configuration: Parameter Data.inputDBS has an invalid value ('%s')." % (
self.configuration.Data.inputDBS)
msg += "\nAllowed values are: "
msg += "\n ".join([
"'%s' ('%s')" % (alias, url)
for alias, url in DBSURLS['reader'].iteritems()
])
local_dbs_urls_aliases = ['phys01', 'phys02', 'phys03']
local_dbs_urls = [
DBSURLS['reader'][alias]
for alias in local_dbs_urls_aliases
if alias in DBSURLS['reader']
]
if self.configuration.Data.inputDBS in local_dbs_urls + local_dbs_urls_aliases:
inputDataset_parts = self.configuration.Data.inputDataset.split(
'/')
inputDataset_parts.pop(0)
inputDataset_tier = inputDataset_parts[-1] if len(
inputDataset_parts) == 3 else None
user_data_tiers = ['USER']
if inputDataset_tier not in user_data_tiers:
msg = "Invalid CRAB configuration: A local DBS instance '%s' was specified for reading an input dataset of tier %s." \
% (self.configuration.Data.inputDBS, inputDataset_tier)
msg += "\nDatasets of tier different than %s must be read from the global DBS instance; this is, set Data.inputDBS = 'global'." \
% (", ".join(user_data_tiers[:-1]) + " or " + user_data_tiers[-1] if len(user_data_tiers) > 1 else user_data_tiers[0])
if msg:
inputDBS_default = getParamDefaultValue('Data.inputDBS')
if inputDBS_default:
inputDBS_default, inputDBS_default_alias = self.getDBSURLAndAlias(
inputDBS_default, 'reader')
if inputDBS_default and inputDBS_default_alias:
msg += "\nIf Data.inputDBS would not be specified, the default '%s' ('%s') would be used." % (
inputDBS_default_alias, inputDBS_default)
return False, msg
## If a publication DBS URL is specified and publication is ON, check that the DBS URL is a good one.
if hasattr(self.configuration.Data, 'publishDBS'):
if getattr(self.configuration.Data, 'publication',
getParamDefaultValue('Data.publication')):
dbs_urls = DBSURLS['writer'].values()
dbs_urls_aliases = DBSURLS['writer'].keys()
if (self.configuration.Data.publishDBS not in dbs_urls_aliases
) and (self.configuration.Data.publishDBS.rstrip('/')
not in dbs_urls):
msg = "Invalid CRAB configuration: Parameter Data.publishDBS has an invalid value ('%s')." % (
self.configuration.Data.publishDBS)
msg += "\nAllowed values are: "
msg += "\n ".join([
"'%s' ('%s')" % (alias, url)
for alias, url in DBSURLS['writer'].iteritems()
])
publishDBS_default = getParamDefaultValue(
'Data.publishDBS')
if publishDBS_default:
publishDBS_default, publishDBS_default_alias = self.getDBSURLAndAlias(
publishDBS_default, 'writer')
if publishDBS_default and publishDBS_default_alias:
msg += "\nIf Data.publishDBS would not be specified, the default '%s' ('%s') would be used." \
% (publishDBS_default_alias, publishDBS_default)
return False, msg
if hasattr(self.configuration.JobType, 'scriptExe'):
if not os.path.isfile(self.configuration.JobType.scriptExe):
msg = "Cannot find the file %s specified in the JobType.scriptExe configuration parameter." % (
self.configuration.JobType.scriptExe)
return False, msg
## If ignoreLocality is set, check that a sitewhilelist is present
if getattr(self.configuration.Data, 'ignoreLocality', False):
if not hasattr(self.configuration.Site, 'whitelist'):
msg = "Invalid CRAB configuration:\n when ignoreLocality is set a valid site white list must be specified using the Site.whitelist parameter"
return False, msg
if hasattr(self.configuration.General, 'failureLimit'):
msg = "You have specified deprecated parameter 'failureLimit' which will be removed in the near future."
msg += "\nIf you really need it write a mail to hn-cms-computingTools explaining your use case."
self.logger.warning("%sWARNING%s: %s" %
(colors.RED, colors.NORMAL, msg))
return True, "Valid configuration"
def __call__(self):
self.logger.debug("Started submission")
serverFactory = CRABClient.Emulator.getEmulator('rest')
uniquerequestname = None
self.logger.debug("Working on %s" % str(self.requestarea))
self.configreq = {'dryrun': 1 if self.options.dryrun else 0}
for param in parametersMapping['on-server']:
mustbetype = getattr(types,
parametersMapping['on-server'][param]['type'])
default = parametersMapping['on-server'][param]['default']
config_params = parametersMapping['on-server'][param]['config']
for config_param in config_params:
attrs = config_param.split('.')
temp = self.configuration
for attr in attrs:
temp = getattr(temp, attr, None)
if temp is None:
break
if temp is not None:
self.configreq[param] = temp
break
elif default is not None:
self.configreq[param] = default
temp = default
else:
## Parameter not strictly required.
pass
## Check that the requestname is of the right type.
## This is not checked in SubCommand.validateConfig().
if param == 'workflow':
if isinstance(self.requestname, mustbetype):
self.configreq['workflow'] = self.requestname
## Translate boolean flags into integers.
elif param in ['savelogsflag', 'publication', 'publishgroupname', 'nonprodsw', 'useparent',\
'ignorelocality', 'saveoutput', 'oneEventMode', 'nonvaliddata', 'ignoreglobalblacklist']:
self.configreq[param] = 1 if temp else 0
## Translate DBS URL aliases into DBS URLs.
elif param in ['dbsurl', 'publishdbsurl']:
if param == 'dbsurl':
dbstype = 'reader'
elif param == 'publishdbsurl':
dbstype = 'writer'
allowed_dbsurls = DBSURLS[dbstype].values()
allowed_dbsurls_aliases = DBSURLS[dbstype].keys()
if self.configreq[param] in allowed_dbsurls_aliases:
self.configreq[param] = DBSURLS[dbstype][
self.configreq[param]]
elif self.configreq[param].rstrip('/') in allowed_dbsurls:
self.configreq[param] = self.configreq[param].rstrip('/')
elif param == 'scriptexe' and 'scriptexe' in self.configreq:
self.configreq[param] = os.path.basename(self.configreq[param])
jobconfig = {}
#get the backend URLs from the server external configuration
serverBackendURLs = server_info('backendurls', self.serverurl,
self.proxyfilename,
getUrl(self.instance, resource='info'))
#if cacheSSL is specified in the server external configuration we will use it to upload the sandbox
filecacheurl = serverBackendURLs[
'cacheSSL'] if 'cacheSSL' in serverBackendURLs else None
pluginParams = [
self.configuration, self.proxyfilename, self.logger,
os.path.join(self.requestarea, 'inputs')
]
crab_job_types = getJobTypes()
if upper(self.configreq['jobtype']) in crab_job_types:
plugjobtype = crab_job_types[upper(
self.configreq['jobtype'])](*pluginParams)
dummy_inputfiles, jobconfig = plugjobtype.run(filecacheurl)
else:
fullname = self.configreq['jobtype']
basename = os.path.basename(fullname).split('.')[0]
plugin = addPlugin(fullname)[basename]
pluginInst = plugin(*pluginParams)
dummy_inputfiles, jobconfig = pluginInst.run()
if self.configreq['publication']:
non_edm_files = jobconfig['tfileoutfiles'] + jobconfig[
'addoutputfiles']
if non_edm_files:
msg = "%sWarning%s: The following output files will not be published, as they are not EDM files: %s" % (
colors.RED, colors.NORMAL, non_edm_files)
self.logger.warning(msg)
self.configreq.update(jobconfig)
server = serverFactory(self.serverurl,
self.proxyfilename,
self.proxyfilename,
version=__version__)
self.logger.info("Sending the request to the server at %s" %
self.serverurl)
self.logger.debug("Submitting %s " % str(self.configreq))
## TODO: this shouldn't be hard-coded.
listParams = ['addoutputfiles', 'sitewhitelist', 'siteblacklist', 'blockwhitelist', 'blockblacklist', \
'tfileoutfiles', 'edmoutfiles', 'runs', 'lumis', 'userfiles', 'scriptargs', 'extrajdl']
self.configreq_encoded = self._encodeRequest(self.configreq,
listParams)
self.logger.debug('Encoded submit request: %s' %
(self.configreq_encoded))
dictresult, status, reason = server.put(self.uri,
data=self.configreq_encoded)
self.logger.debug("Result: %s" % dictresult)
if status != 200:
msg = "Problem sending the request:\ninput:%s\noutput:%s\nreason:%s" % (
str(self.configreq), str(dictresult), str(reason))
raise RESTCommunicationException(msg)
elif 'result' in dictresult:
uniquerequestname = dictresult["result"][0]["RequestName"]
else:
msg = "Problem during submission, no request ID returned:\ninput:%s\noutput:%s\nreason:%s" \
% (str(self.configreq), str(dictresult), str(reason))
raise RESTCommunicationException(msg)
tmpsplit = self.serverurl.split(':')
createCache(self.requestarea,
tmpsplit[0],
tmpsplit[1] if len(tmpsplit) > 1 else '',
uniquerequestname,
voRole=self.voRole,
voGroup=self.voGroup,
instance=self.instance,
originalConfig=self.configuration)
self.logger.info(
"%sSuccess%s: Your task has been delivered to the %s CRAB3 server."
% (colors.GREEN, colors.NORMAL, self.instance))
if not (self.options.wait or self.options.dryrun):
self.logger.info("Task name: %s" % uniquerequestname)
projDir = os.path.join(
getattr(self.configuration.General, 'workArea', '.'),
self.requestname)
self.logger.info("Project dir: %s" % projDir)
self.logger.info(
"Please use 'crab status -d %s' to check how the submission process proceeds.",
projDir)
else:
targetTaskStatus = 'UPLOADED' if self.options.dryrun else 'SUBMITTED'
checkStatusLoop(self.logger, server, self.uri, uniquerequestname,
targetTaskStatus, self.name)
if self.options.dryrun:
self.printDryRunResults(*self.executeTestRun(filecacheurl))
self.logger.debug("About to return")
return {
'requestname': self.requestname,
'uniquerequestname': uniquerequestname
}
def validateConfig(self):
"""
__validateConfig__
Checking if needed input parameters are there
"""
valid, msg = SubCommand.validateConfig(self)
if not valid:
return False, msg
## Check that Data.unitsPerjob is specified.
if hasattr(self.configuration.Data, 'unitsPerJob'):
try:
float(self.configuration.Data.unitsPerJob)
except ValueError:
msg = "Invalid CRAB configuration: Parameter Data.unitsPerJob must be a valid number, not %s." % (
self.configuration.Data.unitsPerJob)
return False, msg
## Check that JobType.pluginName and JobType.externalPluginFile are not both specified.
if hasattr(self.configuration.JobType, 'pluginName') and hasattr(
self.configuration.JobType, 'externalPluginFile'):
msg = "Invalid CRAB configuration: Only one of JobType.pluginName or JobType.externalPluginFile parameters can be specified."
pluginName_default = getParamDefaultValue('JobType.pluginName')
if pluginName_default:
msg += "\nIf neither JobType.pluginName nor JobType.externalPluginFile would be specified,"
msg += " the default JobType.pluginName = '%s' would be used." % (
pluginName_default)
return False, msg
## Load the external plugin or check that the crab plugin is valid.
external_plugin_name = getattr(self.configuration.JobType,
'externalPluginFile', None)
crab_plugin_name = getattr(self.configuration.JobType, 'pluginName',
None)
crab_job_types = {'ANALYSIS': None, 'PRIVATEMC': None} #getJobTypes()
if external_plugin_name:
addPlugin(external_plugin_name) # Do we need to do this here?
if crab_plugin_name:
if upper(crab_plugin_name) not in crab_job_types:
msg = "Invalid CRAB configuration: Parameter JobType.pluginName has an invalid value ('%s')." % (
crab_plugin_name)
msg += "\nAllowed values are: %s." % (", ".join(
['%s' % job_type for job_type in crab_job_types.keys()]))
return False, msg
msg = "Will use CRAB %s plugin" % ("Analysis" if upper(
crab_plugin_name) == 'ANALYSIS' else "PrivateMC")
msg += " (i.e. will run %s job type)." % ("an analysis" if upper(
crab_plugin_name) == 'ANALYSIS' else "a MC generation")
self.logger.debug(msg)
## Check that the particular combination (Data.publication = True, General.transferOutputs = False) is not specified.
if getattr(self.configuration.Data, 'publication', getParamDefaultValue('Data.publication')) and \
not getattr(self.configuration.General, 'transferOutputs', getParamDefaultValue('General.transferOutputs')):
msg = "Invalid CRAB configuration: Data.publication is True, but General.transferOutputs is False."
msg += "\nPublication can not be performed if the output files are not transferred to a permanent storage."
return False, msg
## Check that a storage site is specified if General.transferOutputs = True or General.transferLogs = True.
if not hasattr(self.configuration.Site, 'storageSite'):
if getattr(self.configuration.General, 'transferLogs', getParamDefaultValue('General.transferLogs')) or \
getattr(self.configuration.General, 'transferOutputs', getParamDefaultValue('General.transferOutputs')):
msg = "Invalid CRAB configuration: Parameter Site.storageSite is missing."
return False, msg
## If an input dataset and a DBS URL are specified, check that the DBS URL is a good one.
## Also, if the DBS URL is 'phys0x', check that the input dataset tier is USER.
if hasattr(self.configuration.Data, 'inputDBS'):
if hasattr(self.configuration.Data, 'inputDataset'):
msg = None
dbs_urls_aliases = DBSURLS['reader'].keys()
dbs_urls = DBSURLS['reader'].values()
if (self.configuration.Data.inputDBS not in dbs_urls_aliases
) and (self.configuration.Data.inputDBS.rstrip('/')
not in dbs_urls):
msg = "Invalid CRAB configuration: Parameter Data.inputDBS has an invalid value ('%s')." % (
self.configuration.Data.inputDBS)
msg += "\nAllowed values are: "
msg += "\n ".join([
"'%s' ('%s')" % (alias, url)
for alias, url in DBSURLS['reader'].iteritems()
])
local_dbs_urls_aliases = ['phys01', 'phys02', 'phys03']
local_dbs_urls = [
DBSURLS['reader'][alias]
for alias in local_dbs_urls_aliases
if alias in DBSURLS['reader']
]
if self.configuration.Data.inputDBS in local_dbs_urls + local_dbs_urls_aliases:
inputDataset_parts = self.configuration.Data.inputDataset.split(
'/')
inputDataset_parts.pop(0)
inputDataset_tier = inputDataset_parts[-1] if len(
inputDataset_parts) == 3 else None
user_data_tiers = ['USER']
if inputDataset_tier not in user_data_tiers:
msg = "Invalid CRAB configuration: A local DBS instance '%s' was specified for reading an input dataset of tier %s." \
% (self.configuration.Data.inputDBS, inputDataset_tier)
msg += "\nDatasets of tier different than %s must be read from the global DBS instance; this is, set Data.inputDBS = 'global'." \
% (", ".join(user_data_tiers[:-1]) + " or " + user_data_tiers[-1] if len(user_data_tiers) > 1 else user_data_tiers[0])
if msg:
inputDBS_default = getParamDefaultValue('Data.inputDBS')
if inputDBS_default:
inputDBS_default, inputDBS_default_alias = self.getDBSURLAndAlias(
inputDBS_default, 'reader')
if inputDBS_default and inputDBS_default_alias:
msg += "\nIf Data.inputDBS would not be specified, the default '%s' ('%s') would be used." % (
inputDBS_default_alias, inputDBS_default)
return False, msg
## If a publication DBS URL is specified and publication is ON, check that the DBS URL is a good one.
if hasattr(self.configuration.Data, 'publishDBS'):
if getattr(self.configuration.Data, 'publication',
getParamDefaultValue('Data.publication')):
dbs_urls = DBSURLS['writer'].values()
dbs_urls_aliases = DBSURLS['writer'].keys()
if (self.configuration.Data.publishDBS not in dbs_urls_aliases
) and (self.configuration.Data.publishDBS.rstrip('/')
not in dbs_urls):
msg = "Invalid CRAB configuration: Parameter Data.publishDBS has an invalid value ('%s')." % (
self.configuration.Data.publishDBS)
msg += "\nAllowed values are: "
msg += "\n ".join([
"'%s' ('%s')" % (alias, url)
for alias, url in DBSURLS['writer'].iteritems()
])
publishDBS_default = getParamDefaultValue(
'Data.publishDBS')
if publishDBS_default:
publishDBS_default, publishDBS_default_alias = self.getDBSURLAndAlias(
publishDBS_default, 'writer')
if publishDBS_default and publishDBS_default_alias:
msg += "\nIf Data.publishDBS would not be specified, the default '%s' ('%s') would be used." \
% (publishDBS_default_alias, publishDBS_default)
return False, msg
if hasattr(self.configuration.JobType, 'scriptExe'):
if not os.path.isfile(self.configuration.JobType.scriptExe):
msg = "Cannot find the file %s specified in the JobType.scriptExe configuration parameter." % (
self.configuration.JobType.scriptExe)
return False, msg
return True, "Valid configuration"
