def closeOutStep0Requests(url, workflows):
"""
Closes either montecarlo step0 requests
"""
noSiteWorkflows = []
for workflow in workflows:
datasets = reqMgrClient.outputdatasetsWorkflow(url, workflow)
status = reqMgrClient.getWorkflowStatus(url, workflow)
#if not completed skip
if status != 'completed':
continue
closeOutWorkflow = True
#skip montecarlos on a special queue
if reqMgrClient.getRequestTeam(url, workflow) == 'analysis':
continue
for dataset in datasets:
closeOutDataset = False
percentage = percentageCompletion(url, workflow, dataset)
phedexSubscription = phedexClient.getCustodialMoveSubscriptionSite(dataset)
transPerc = 0
closedBlocks = None
duplicate = None
correctLumis = None
# if dataset has subscription and enough events we check
# duplicates, transfer percentage, closed blocks and lumis
if phedexSubscription and percentage >= float(0.95):
transPerc = phedexClient.getTransferPercentage(url, dataset, phedexSubscription)
duplicate = dbs3Client.duplicateLumi(dataset)
correctLumis = checkCorrectLumisEventGEN(dataset)
#TODO validate closed blocks
if not duplicate and correctLumis:
closeOutDataset = True
else:
closeOutDataset = False
#validate when percentage is ok but has not phedex subscription
if percentage >= float(0.95) and not phedexSubscription:
noSiteWorkflows.append(workflow)
#if at least one dataset is not ready wf cannot be closed out
closeOutWorkflow = closeOutWorkflow and closeOutDataset
print '| %80s | %100s | %4s | %5s| %3s | %5s| %5s| ' % (workflow, dataset,str(int(percentage*100)),
str(phedexSubscription), str(correctLumis), duplicate, closeOutDataset)
#workflow can only be closed out if all datasets are ready
if closeOutWorkflow:
reqMgrClient.closeOutWorkflowCascade(url, workflow)
print '-'*180
return noSiteWorkflows
def closeOutMonterCarloRequests(url, workflows):
"""
Closes either montecarlo or montecarlo from gen
workflows
"""
noSiteWorkflows = []
for workflow in workflows:
datasets = reqMgrClient.outputdatasetsWorkflow(url, workflow)
closeOutWorkflow = True
#skip montecarlos on a special queue
if reqMgrClient.getRequestTeam(url, workflow) == 'analysis':
continue
for dataset in datasets:
closePercentage = 0.95
# validation for SMS montecarlos
if 'SMS' in dataset:
closePercentage= 1.00
percentage = percentageCompletion(url, workflow, dataset)
phedexSubscription = phedexClient.getCustodialMoveSubscriptionSite(dataset)
transPerc = 0
closedBlocks = None
duplicate = None
# if dataset has subscription and enough events we check
# duplicates, transfer percentage and closed blocks
if phedexSubscription and percentage >= float(closePercentage):
transPerc = phedexClient.getTransferPercentage(url, dataset, phedexSubscription)
duplicate = dbs3Client.duplicateLumi(dataset)
if not duplicate:
closeOutDataset = True
else:
closeOutDataset = False
else:
closeOutDataset = False
#validate when percentage is ok but has not phedex subscription
if percentage >= float(closePercentage) and not phedexSubscription:
noSiteWorkflows.append(workflow)
#if at least one dataset is not ready wf cannot be closed out
closeOutWorkflow = closeOutWorkflow and closeOutDataset
print '| %80s | %100s | %4s | %5s| %3s | %5s| %5s|' % (workflow, dataset,str(int(percentage*100)),
str(phedexSubscription), str(int(transPerc*100)), duplicate, closeOutDataset)
#workflow can only be closed out if all datasets are ready
if closeOutWorkflow:
reqMgrClient.closeOutWorkflowCascade(url, workflow)
#separation line
print '-'*180
return noSiteWorkflows
def validateClosingTaskChain(url, workflow):
"""
Calculates a Percentage completion for a taskchain.
Taking step/filter efficiency into account.
test with pdmvserv_task_SUS-Summer12WMLHE-00004__v1_T_141003_120119_9755
"""
inputEvents = workflow.getInputEvents()
#if subtype doesn't come with the request, we decide based on dataset names
fromGen = False
#if no output dataset ends with GEN or LHE
if not re.match('.*/(GEN|LHE)$', workflow.outputDatasets[0]):
fromGen = False
elif (re.match('.*/(GEN|LHE)$', workflow.outputDatasets[0])
and re.match('.*/(GEN-SIM|GEN)$', workflow.outputDatasets[1])):
fromGen = True
#task-chain 1 (without filterEff)
if not fromGen:
#validate with the regular procedure
result = validateClosingWorkflow(url, workflow, 0.95)
return result
#task-chain 2 GEN, GEN-SIM, GEN-SIM-RAW, AODSIM, DQM
else:
#GEN/LHE and GEN-SIM
result = {'name':workflow.name, 'datasets': {}}
result['datasets'] = dict( (ds,{}) for ds in workflow.outputDatasets)
closeOutWorkflow = True
i = 1
for dataset in workflow.outputDatasets:
closeOutDataset = False
#percentage
outputEvents = workflow.getOutputEvents(dataset)
filterEff = workflow.getFilterEfficiency('Task%d'%i)
#GEN/LHE and GEN-SIM
if 1 <= i <= 2:
#decrease filter eff
inputEvents *= filterEff
#percentage
percentage = outputEvents/float(inputEvents) if inputEvents > 0 else 0.0
#phedex request
phedexReqs = phedexClient.getCustodialSubscriptionRequestSite(dataset)
#all validations
duplicate = None
correctLumis = None
transPerc = None
missingSubs = False
#TODO test
dbsFiles = dbs3Client.getFileCountDataset(dataset)
phdFiles = phedexClient.getFileCountDataset(url,dataset)
equalFiles = (dbsFiles == phdFiles)
#Check first percentage
if percentage >= 0.95:
#if we need to check duplicates
duplicate = dbs3Client.duplicateRunLumi(dataset)
#dataset healthy means:
# checkDuplicates -> no duplicates
if not duplicate:
#if phedex check not required we can closeout
if phedexReqs:
try:
transPerc = phedexClient.getTransferPercentage(url, dataset, phedexReqs[0])
except:
transPerc = None
#last check if files are equal
closeOutDataset = equalFiles
else:
missingSubs = True
#if at least one dataset is not ready wf cannot be closed out
closeOutWorkflow = closeOutWorkflow and closeOutDataset
#load results in a dict
result['datasets'][dataset]["percentage"] = percentage
result['datasets'][dataset]["duplicate"] = duplicate
result['datasets'][dataset]["phedexReqs"] = phedexReqs
result['datasets'][dataset]["closeOutDataset"] = closeOutDataset
result['datasets'][dataset]["transPerc"] = transPerc
result['datasets'][dataset]["correctLumis"] = correctLumis
result['datasets'][dataset]["missingSubs"] = missingSubs
result['datasets'][dataset]["dbsFiles"] = dbsFiles
result['datasets'][dataset]["phedexFiles"] = phdFiles
i += 1
result['closeOutWorkflow'] = closeOutWorkflow
return result
def validateClosingWorkflow(url, workflow, closePercentage = 0.95, checkEqual=False,
checkDuplicates=True, checkLumiNumb=False, checkPhedex='custodial'):
"""
Validates if a workflow can be closed out, using different parameters of validation.
returns the response as a dict.
checkPhedex can be 'custodial', 'any' or False
"""
#inputDataset = reqMgrClient.getInputDataSet(url, workflow)
result = {'name':workflow.name, 'datasets': {}}
result['datasets'] = dict( (ds,{}) for ds in workflow.outputDatasets)
closeOutWorkflow = True
#check if dataset is ready
#TODO validate here if workflow is MonteCarlo from GEN with two output
for dataset in workflow.outputDatasets:
closeOutDataset = False
try:
percentage = workflow.percentageCompletion(dataset, skipInvalid=True)
except Exception as e:
print 'Error getting information from DBS', workflow, dataset
percentage = 0.0
#retrieve either custodial or all subscriptions.
try:
if checkPhedex == 'custodial':
phedexReqs = phedexClient.getCustodialSubscriptionRequestSite(dataset)
elif checkPhedex == 'any':
phedexReqs = phedexClient.getSubscriptionSites(dataset)
else:
phedexReqs = None
except Exception:
print 'Error getting phedex info,: ', dataset
phedexReqs = None
duplicate = None
correctLumis = None
transPerc = None
missingSubs = False
equalFiles = None
dbsFiles = dbs3Client.getFileCountDataset(dataset)
phdFiles = phedexClient.getFileCountDataset(url,dataset)
equalFiles = (dbsFiles == phdFiles)
#Check first percentage
if ((checkEqual and percentage == closePercentage)
or (not checkEqual and percentage >= closePercentage)
or dataset.endswith("DQMIO") ): #DQMIO are exceptions (have 0 events)
#if we need to check duplicates
if checkDuplicates:
try:
duplicate = dbs3Client.duplicateRunLumi(dataset, skipInvalid=True)
except Exception:
print "Error in checking duplicate lumis for", dataset
#if we need to check for correct lumi number
if checkLumiNumb:
correctLumis = checkCorrectLumisEventGEN(dataset)
#dataset healthy means:
# checkDuplicates -> no duplicates
# checkLumiNumb -> correct
if (not (checkDuplicates and duplicate) and
not ( checkLumiNumb and not correctLumis)):
#if phedex check not required we can closeout
if not checkPhedex:
#last check, that files are equal
closeOutDataset = equalFiles
#if phedex check is required and has it
elif checkPhedex and phedexReqs:
try:
transPerc = phedexClient.getTransferPercentage(url, dataset, phedexReqs[0])
except:
transPerc = None
#last check, that files are equal
closeOutDataset = equalFiles
else:
#TODO only missing subscription if equal # of files
missingSubs = equalFiles
#if at least one dataset is not ready wf cannot be closed out
closeOutWorkflow = closeOutWorkflow and closeOutDataset
#load results in a dict
result['datasets'][dataset]["percentage"] = percentage
result['datasets'][dataset]["duplicate"] = duplicate
result['datasets'][dataset]["phedexReqs"] = phedexReqs
result['datasets'][dataset]["closeOutDataset"] = closeOutDataset
result['datasets'][dataset]["transPerc"] = transPerc
result['datasets'][dataset]["correctLumis"] = correctLumis
result['datasets'][dataset]["missingSubs"] = missingSubs
result['datasets'][dataset]["dbsFiles"] = dbsFiles
result['datasets'][dataset]["phedexFiles"] = phdFiles
result['closeOutWorkflow'] = closeOutWorkflow
return result
