def testCreateDeleteExistsNoFiles(self):
"""
_testCreateDeleteExistsNoFiles_
Create and then delete a job but don't add any input files to it.
Use the job class's exists() method to determine if the job has been
written to the database before it is created, after it has been created
and after it has been deleted.
"""
testWorkflow = Workflow(spec="spec.xml", owner="Simon", name="wf001", task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset, workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testJob = Job(name="TestJob")
assert testJob.exists() == False, "ERROR: Job exists before it was created"
testJob.create(group=testJobGroup)
assert testJob.exists() >= 0, "ERROR: Job does not exist after it was created"
testJob.delete()
assert testJob.exists() == False, "ERROR: Job exists after it was delete"
return
def createTestJob(subscriptionType="Merge"):
"""
_createTestJob_
Create a test job with two files as input. This will also create the
appropriate workflow, jobgroup and subscription.
"""
testWorkflow = Workflow(spec=makeUUID(), owner="Simon", name=makeUUID(), task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset, workflow=testWorkflow, type=subscriptionType)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileA.addRun(Run(1, *[45]))
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileB.addRun(Run(1, *[46]))
testFileA.create()
testFileB.create()
testJob = Job(name=makeUUID(), files=[testFileA, testFileB])
testJob["couch_record"] = "somecouchrecord"
testJob["location"] = "test.site.ch"
testJob.create(group=testJobGroup)
testJob.associateFiles()
return testJob
def testMask(self):
"""
_testMask_
Test the new mask setup
"""
testWorkflow = Workflow(spec="spec.xml", owner="Steve",
name="wf001", task="Test")
testWorkflow.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow)
testSubscription.create()
testFileA = File(lfn=makeUUID(), locations="test.site.ch")
testFileB = File(lfn=makeUUID(), locations="test.site.ch")
testFileA.create()
testFileB.create()
testFileset.addFile([testFileA, testFileB])
testFileset.commit()
testSubscription.acquireFiles([testFileA, testFileB])
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testJob = Job()
testJob['mask'].addRunAndLumis(run=100, lumis=[101, 102])
testJob['mask'].addRunAndLumis(run=200, lumis=[201, 202])
testJob.create(group=testJobGroup)
loadJob = Job(id=testJob.exists())
loadJob.loadData()
runs = loadJob['mask'].getRunAndLumis()
self.assertEqual(len(runs), 2)
self.assertEqual(runs[100], [[101, 102]])
self.assertEqual(runs[200], [[201, 202]])
bigRun = Run(100, *[101, 102, 103, 104])
badRun = Run(300, *[1001, 1002])
result = loadJob['mask'].filterRunLumisByMask([bigRun, badRun])
self.assertEqual(len(result), 1)
alteredRun = result.pop()
self.assertEqual(alteredRun.run, 100)
self.assertEqual(alteredRun.lumis, [101, 102])
run0 = Run(300, *[1001, 1002])
run1 = Run(300, *[1001, 1002])
loadJob['mask'].filterRunLumisByMask([run0, run1])
return
def testCreateDeleteExists(self):
"""
_testCreateDeleteExists_
Create and then delete a job. Use the job class's exists() method to
determine if the job has been written to the database before it is
created, after it has been created and after it has been deleted.
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Simon",
name="wf001",
task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileA.create()
testFileB.create()
testJob = Job(name="TestJob", files=[testFileA, testFileB])
testWU = WorkUnit(taskID=testWorkflow.id,
fileid=testFileA['id'],
runLumi=Run(1, *[44]))
self.assertFalse(testJob.exists(), "Job exists before it was created")
self.assertFalse(testWU.exists(),
"WorkUnit exists before it was created")
testJob.create(group=testJobGroup)
self.assertTrue(testJob.exists(),
"Job does not exist after it was created")
self.assertFalse(testWU.exists(),
"WorkUnit exists when there is no work")
# Test the getWorkflow method
workflow = testJob.getWorkflow()
self.assertEqual(workflow['task'], 'Test')
self.assertEqual(workflow['name'], 'wf001')
testJob.delete()
self.assertFalse(testJob.exists(), "Job exists after it was deleted")
self.assertFalse(testWU.exists(),
"WorkUnit exists after job is deleted")
return
def makeNJobs(self, name, task, nJobs, jobGroup, fileset, sub, site = None, bl = [], wl = []):
"""
_makeNJobs_
Make and return a WMBS Job and File
This handles all those damn add-ons
"""
# Set the CacheDir
cacheDir = os.path.join(self.testDir, 'CacheDir')
for n in range(nJobs):
# First make a file
#site = self.sites[0]
testFile = File(lfn = "/singleLfn/%s/%s" %(name, n),
size = 1024, events = 10)
if site:
testFile.setLocation(site)
else:
for tmpSite in self.sites:
testFile.setLocation('se.%s' % (tmpSite))
testFile.create()
fileset.addFile(testFile)
fileset.commit()
index = 0
for f in fileset.files:
index += 1
testJob = Job(name = '%s-%i' %(name, index))
testJob.addFile(f)
testJob["location"] = f.getLocations()[0]
testJob['custom']['location'] = f.getLocations()[0]
testJob['task'] = task.getPathName()
testJob['sandbox'] = task.data.input.sandbox
testJob['spec'] = os.path.join(self.testDir, 'basicWorkload.pcl')
testJob['mask']['FirstEvent'] = 101
testJob['owner'] = 'tapas'
testJob["siteBlacklist"] = bl
testJob["siteWhitelist"] = wl
testJob['ownerDN'] = 'tapas'
testJob['ownerRole'] = 'cmsrole'
testJob['ownerGroup'] = 'phgroup'
jobCache = os.path.join(cacheDir, 'Sub_%i' % (sub), 'Job_%i' % (index))
os.makedirs(jobCache)
testJob.create(jobGroup)
testJob['cache_dir'] = jobCache
testJob.save()
jobGroup.add(testJob)
output = open(os.path.join(jobCache, 'job.pkl'),'w')
pickle.dump(testJob, output)
output.close()
return testJob, testFile
def testDeleteTransaction(self):
"""
_testDeleteTransaction_
Create a new job and commit it to the database. Start a new transaction
and delete the file from the database. Verify that the file has been
deleted. After that, roll back the transaction and verify that the
job is once again in the database.
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Simon",
name="wf001",
task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileA.create()
testFileB.create()
testJob = Job(name="TestJob", files=[testFileA, testFileB])
assert testJob.exists() == False, \
"ERROR: Job exists before it was created"
testJob.create(group=testJobGroup)
assert testJob.exists() >= 0, \
"ERROR: Job does not exist after it was created"
myThread = threading.currentThread()
myThread.transaction.begin()
testJob.delete()
assert testJob.exists() == False, \
"ERROR: Job exists after it was delete"
myThread.transaction.rollback()
assert testJob.exists() >= 0, \
"ERROR: Job does not exist after transaction was rolled back."
return
def makeNJobs(self, name, task, nJobs, jobGroup, fileset, sub, site=None, bl=[], wl=[]):
"""
_makeNJobs_
Make and return a WMBS Job and File
This handles all those damn add-ons
"""
# Set the CacheDir
cacheDir = os.path.join(self.testDir, "CacheDir")
for n in range(nJobs):
# First make a file
# site = self.sites[0]
testFile = File(lfn="/singleLfn/%s/%s" % (name, n), size=1024, events=10)
if site:
testFile.setLocation(site)
else:
for tmpSite in self.sites:
testFile.setLocation("se.%s" % (tmpSite))
testFile.create()
fileset.addFile(testFile)
fileset.commit()
index = 0
for f in fileset.files:
index += 1
testJob = Job(name="%s-%i" % (name, index))
testJob.addFile(f)
testJob["location"] = f.getLocations()[0]
testJob["custom"]["location"] = f.getLocations()[0]
testJob["task"] = task.getPathName()
testJob["sandbox"] = task.data.input.sandbox
testJob["spec"] = os.path.join(self.testDir, "basicWorkload.pcl")
testJob["mask"]["FirstEvent"] = 101
testJob["owner"] = "tapas"
testJob["siteBlacklist"] = bl
testJob["siteWhitelist"] = wl
testJob["ownerDN"] = "tapas"
testJob["ownerRole"] = "cmsrole"
testJob["ownerGroup"] = "phgroup"
jobCache = os.path.join(cacheDir, "Sub_%i" % (sub), "Job_%i" % (index))
os.makedirs(jobCache)
testJob.create(jobGroup)
testJob["cache_dir"] = jobCache
testJob.save()
jobGroup.add(testJob)
output = open(os.path.join(jobCache, "job.pkl"), "w")
pickle.dump(testJob, output)
output.close()
return testJob, testFile
def makeNJobs(self, name, task, nJobs, jobGroup, fileset, sub, site):
"""
_makeNJobs_
Make and return a WMBS Job and File
This handles all those damn add-ons
"""
# Set the CacheDir
cacheDir = os.path.join(self.testDir, 'CacheDir')
for n in range(nJobs):
# First make a file
# site = self.sites[0]
testFile = File(lfn="/singleLfn/%s/%s" % (name, n),
size=1024,
events=10)
fileset.addFile(testFile)
fileset.commit()
location = None
if isinstance(site, list):
if len(site) > 0:
location = site[0]
else:
location = site
index = 0
for f in fileset.files:
index += 1
testJob = Job(name='%s-%i' % (name, index))
testJob.addFile(f)
testJob["location"] = location
testJob["possiblePSN"] = set(site) if isinstance(
site, list) else set([site])
testJob['task'] = task.getPathName()
testJob['sandbox'] = task.data.input.sandbox
testJob['spec'] = os.path.join(self.testDir, 'basicWorkload.pcl')
testJob['mask']['FirstEvent'] = 101
testJob['priority'] = 101
testJob['numberOfCores'] = 1
jobCache = os.path.join(cacheDir, 'Sub_%i' % (sub),
'Job_%i' % (index))
os.makedirs(jobCache)
testJob.create(jobGroup)
testJob['cache_dir'] = jobCache
testJob.save()
jobGroup.add(testJob)
output = open(os.path.join(jobCache, 'job.pkl'), 'w')
pickle.dump(testJob, output)
output.close()
return testJob, testFile
def testParentageByJob(self):
"""
_testParentageByJob_
Tests the DAO that assigns parentage by Job
"""
testWorkflow = Workflow(spec = 'hello', owner = "mnorman",
name = "wf001", task="basicWorkload/Production")
testWorkflow.create()
testFileset = Fileset(name = "TestFileset")
testFileset.create()
testSubscription = Subscription(fileset = testFileset, workflow = testWorkflow, type = "Processing", split_algo = "FileBased")
testSubscription.create()
testJobGroup = JobGroup(subscription = testSubscription)
testJobGroup.create()
testFileParentA = File(lfn = "/this/is/a/parent/lfnA", size = 1024,
events = 20, checksums = {'cksum': 1})
testFileParentA.addRun(Run( 1, *[45]))
testFileParentB = File(lfn = "/this/is/a/parent/lfnB", size = 1024,
events = 20, checksums = {'cksum': 1})
testFileParentB.addRun(Run( 1, *[45]))
testFileParentA.create()
testFileParentB.create()
testFileA = File(lfn = "/this/is/a/lfn", size = 1024, events = 10,
checksums = {'cksum':1})
testFileA.addRun(Run( 1, *[45]))
testFileA.create()
testJobA = Job()
testJobA.create(group = testJobGroup)
testJobA.addFile(testFileParentA)
testJobA.addFile(testFileParentB)
testJobA.associateFiles()
parentAction = self.daofactory(classname = "Files.SetParentageByJob")
parentAction.execute(binds = {'jobid': testJobA.exists(), 'child': testFileA['lfn']})
testFileB = File(id = testFileA["id"])
testFileB.loadData(parentage = 1)
goldenFiles = [testFileParentA, testFileParentB]
for parentFile in testFileB["parents"]:
self.assertEqual(parentFile in goldenFiles, True,
"ERROR: Unknown parent file")
goldenFiles.remove(parentFile)
self.assertEqual(len(goldenFiles), 0,
"ERROR: Some parents are missing")
def makeNJobs(self, name, task, nJobs, jobGroup, fileset, sub, site, bl = [], wl = []):
"""
_makeNJobs_
Make and return a WMBS Job and File
This handles all those damn add-ons
"""
# Set the CacheDir
cacheDir = os.path.join(self.testDir, 'CacheDir')
for n in range(nJobs):
# First make a file
#site = self.sites[0]
testFile = File(lfn = "/singleLfn/%s/%s" % (name, n),
size = 1024, events = 10)
if type(site) == list:
for singleSite in site:
testFile.setLocation(singleSite)
else:
testFile.setLocation(site)
testFile.create()
fileset.addFile(testFile)
fileset.commit()
index = 0
for f in fileset.files:
index += 1
testJob = Job(name = '%s-%i' % (name, index))
testJob.addFile(f)
testJob["location"] = f.getLocations()[0]
testJob['task'] = task.getPathName()
testJob['sandbox'] = task.data.input.sandbox
testJob['spec'] = os.path.join(self.testDir, 'basicWorkload.pcl')
testJob['mask']['FirstEvent'] = 101
testJob["siteBlacklist"] = bl
testJob["siteWhitelist"] = wl
testJob['priority'] = 101
testJob['multicoreEnabled'] = False
testJob['numberOfCores'] = 1
jobCache = os.path.join(cacheDir, 'Sub_%i' % (sub), 'Job_%i' % (index))
os.makedirs(jobCache)
testJob.create(jobGroup)
testJob['cache_dir'] = jobCache
testJob.save()
jobGroup.add(testJob)
output = open(os.path.join(jobCache, 'job.pkl'), 'w')
pickle.dump(testJob, output)
output.close()
return testJob, testFile
def testDeleteTransaction(self):
"""
_testDeleteTransaction_
Create a new job and commit it to the database. Start a new transaction
and delete the file from the database. Verify that the file has been
deleted. After that, roll back the transaction and verify that the
job is once again in the database.
"""
testWorkflow = Workflow(spec="spec.xml", owner="Simon",
name="wf001", task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileA.create()
testFileB.create()
testJob = Job(name="TestJob", files=[testFileA, testFileB])
assert testJob.exists() is False, \
"ERROR: Job exists before it was created"
testJob.create(group=testJobGroup)
assert testJob.exists() >= 0, \
"ERROR: Job does not exist after it was created"
myThread = threading.currentThread()
myThread.transaction.begin()
testJob.delete()
assert testJob.exists() is False, \
"ERROR: Job exists after it was delete"
myThread.transaction.rollback()
assert testJob.exists() >= 0, \
"ERROR: Job does not exist after transaction was rolled back."
return
def makeNJobs(self, name, task, nJobs, jobGroup, fileset, sub, site):
"""
_makeNJobs_
Make and return a WMBS Job and File
This handles all those damn add-ons
"""
# Set the CacheDir
cacheDir = os.path.join(self.testDir, 'CacheDir')
for n in range(nJobs):
# First make a file
#site = self.sites[0]
testFile = File(lfn = "/singleLfn/%s/%s" % (name, n),
size = 1024, events = 10)
fileset.addFile(testFile)
fileset.commit()
location = None
if isinstance(site, list):
if len(site) > 0:
location = site[0]
else:
location = site
index = 0
for f in fileset.files:
index += 1
testJob = Job(name = '%s-%i' % (name, index))
testJob.addFile(f)
testJob["location"] = location
testJob["possiblePSN"] = set(site) if isinstance(site, list) else set([site])
testJob['task'] = task.getPathName()
testJob['sandbox'] = task.data.input.sandbox
testJob['spec'] = os.path.join(self.testDir, 'basicWorkload.pcl')
testJob['mask']['FirstEvent'] = 101
testJob['priority'] = 101
testJob['numberOfCores'] = 1
jobCache = os.path.join(cacheDir, 'Sub_%i' % (sub), 'Job_%i' % (index))
os.makedirs(jobCache)
testJob.create(jobGroup)
testJob['cache_dir'] = jobCache
testJob.save()
jobGroup.add(testJob)
output = open(os.path.join(jobCache, 'job.pkl'), 'w')
pickle.dump(testJob, output)
output.close()
return testJob, testFile
def createTestJobs(self, nJobs, cacheDir):
"""
_createTestJobs_
Create several jobs
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Simon",
name="wf001",
task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow,
type="Processing",
split_algo="FileBased")
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
# Create a file
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileA.addRun(Run(10, *[12312]))
testFileA.setLocation('malpaquet')
testFileA.create()
baseName = makeUUID()
# Now create a job
for i in range(nJobs):
testJob = Job(name='%s-%i' % (baseName, i))
testJob.addFile(testFileA)
testJob['location'] = 'malpaquet'
testJob['retry_count'] = 1
testJob['retry_max'] = 10
testJob.create(testJobGroup)
testJob.save()
testJobGroup.add(testJob)
testJobGroup.commit()
# Set test job caches
for job in testJobGroup.jobs:
job.setCache(cacheDir)
return testJobGroup
def createTestJobs(self, nJobs, cacheDir):
"""
_createTestJobs_
Create several jobs
"""
testWorkflow = Workflow(spec = "spec.xml", owner = "Simon",
name = "wf001", task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name = "TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset = testWMBSFileset,
workflow = testWorkflow,
type = "Processing",
split_algo = "FileBased")
testSubscription.create()
testJobGroup = JobGroup(subscription = testSubscription)
testJobGroup.create()
# Create a file
testFileA = File(lfn = "/this/is/a/lfnA", size = 1024, events = 10)
testFileA.addRun(Run(10, *[12312]))
testFileA.setLocation('malpaquet')
testFileA.create()
baseName = makeUUID()
# Now create a job
for i in range(nJobs):
testJob = Job(name = '%s-%i' % (baseName, i))
testJob.addFile(testFileA)
testJob['location'] = 'malpaquet'
testJob['retry_count'] = 1
testJob['retry_max'] = 10
testJob.create(testJobGroup)
testJob.save()
testJobGroup.add(testJob)
testJobGroup.commit()
# Set test job caches
for job in testJobGroup.jobs:
job.setCache(cacheDir)
return testJobGroup
def test_AutoIncrementCheck(self):
"""
_AutoIncrementCheck_
Test and see whether we can find and set the auto_increment values
"""
myThread = threading.currentThread()
if not myThread.dialect.lower() == "mysql":
return
testWorkflow = Workflow(spec="spec.xml", owner="Steve", name="wf001", task="Test")
testWorkflow.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testSubscription = Subscription(fileset=testFileset, workflow=testWorkflow)
testSubscription.create()
testFileA = File(lfn=makeUUID(), locations="test.site.ch")
testFileB = File(lfn=makeUUID(), locations="test.site.ch")
testFileA.create()
testFileB.create()
testFileset.addFile([testFileA, testFileB])
testFileset.commit()
testSubscription.acquireFiles([testFileA, testFileB])
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
incrementDAO = self.daoFactory(classname="Jobs.AutoIncrementCheck")
incrementDAO.execute()
testJob = Job()
testJob.create(group=testJobGroup)
self.assertEqual(testJob.exists(), 1)
incrementDAO.execute()
testJob = Job()
testJob.create(group=testJobGroup)
self.assertEqual(testJob.exists(), 2)
incrementDAO.execute(input=10)
testJob = Job()
testJob.create(group=testJobGroup)
self.assertEqual(testJob.exists(), 11)
incrementDAO.execute(input=5)
testJob = Job()
testJob.create(group=testJobGroup)
self.assertEqual(testJob.exists(), 12)
return
def notestCreateDeleteExists(self):
"""
Create and then delete a job and workflow. Use the workunit class's exists() method to
determine if the workunit has been written to the database before the job is
created, after the job has been created, and after the workflow has been deleted.
"""
testWorkflow = Workflow(spec="spec.xml", owner="Simon", name="wf001", task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset, workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileA.addRun(Run(1, *[45]))
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileB.addRun(Run(1, *[46]))
testFileA.create()
testFileB.create()
testJob = Job(name="TestJob", files=[testFileA, testFileB])
testWU1 = WorkUnit(taskID=testWorkflow.id, fileid=testFileA['id'], runLumi=Run(1, *[45]))
testWU2 = WorkUnit(taskID=testWorkflow.id, fileid=testFileB['id'], runLumi=Run(1, *[46]))
self.assertFalse(testWU1.exists(), "WorkUnit exists before job was created")
self.assertFalse(testWU2.exists(), "WorkUnit exists before job was created")
testJob.create(group=testJobGroup)
self.assertTrue(testWU1.exists(), "WorkUnit does not exist after job was created")
self.assertTrue(testWU2.exists(), "WorkUnit does not exist after job was created")
testJob.delete()
self.assertTrue(testWU1.exists(), "WorkUnit does not exist after job is deleted")
self.assertTrue(testWU2.exists(), "WorkUnit does not exist after job is deleted")
testWorkflow.delete()
self.assertFalse(testWU1.exists(), "WorkUnit exists after workflow is deleted")
self.assertFalse(testWU2.exists(), "WorkUnit exists after workflow is deleted")
return
def testCreateDeleteExists(self):
"""
_testCreateDeleteExists_
Create and then delete a job. Use the job class's exists() method to
determine if the job has been written to the database before it is
created, after it has been created and after it has been deleted.
"""
testWorkflow = Workflow(spec="spec.xml", owner="Simon",
name="wf001", task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileA.create()
testFileB.create()
testJob = Job(name="TestJob", files=[testFileA, testFileB])
# testWU = WorkUnit(taskID=testWorkflow.id, fileid=testFileA['id'], runLumi=Run(1, *[44]))
self.assertFalse(testJob.exists(), "Job exists before it was created")
# self.assertFalse(testWU.exists(), "WorkUnit exists before it was created")
testJob.create(group=testJobGroup)
self.assertTrue(testJob.exists(), "Job does not exist after it was created")
# self.assertFalse(testWU.exists(), "WorkUnit exists when there is no work")
# Test the getWorkflow method
workflow = testJob.getWorkflow()
self.assertEqual(workflow['task'], 'Test')
self.assertEqual(workflow['name'], 'wf001')
testJob.delete()
self.assertFalse(testJob.exists(), "Job exists after it was deleted")
# self.assertFalse(testWU.exists(), "WorkUnit exists after job is deleted")
return
def createDummyJobs(self, nJobs, location=None):
"""
_createDummyJobs_
Create some dummy jobs
"""
if not location:
location = self.sites[0]
nameStr = makeUUID()
testWorkflow = Workflow(
spec=nameStr,
owner="tapas",
name=nameStr,
task="basicWorkload/Production",
owner_vogroup="phgroup",
owner_vorole="cmsrole",
)
testWorkflow.create()
testFileset = Fileset(name=nameStr)
testFileset.create()
testSubscription = Subscription(
fileset=testFileset, workflow=testWorkflow, type="Processing", split_algo="FileBased"
)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
jobList = []
for i in range(nJobs):
testJob = Job(name="%s-%i" % (nameStr, i))
testJob["location"] = location
testJob["custom"]["location"] = location
testJob["userdn"] = "tapas"
testJob["owner"] = "tapas"
testJob["userrole"] = "cmsrole"
testJob["usergroup"] = "phgroup"
testJob.create(testJobGroup)
jobList.append(testJob)
return jobList
def createTestJob(self, testSubscription, jobName, *testFiles):
"""
_createTestJob_
Create a test job with two files as input. This will also create the
appropriate workflow, jobgroup and subscription.
"""
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFiles = list(testFiles)
testJob = Job(name=jobName, files=testFiles)
testJob["couch_record"] = "somecouchrecord"
testJob["location"] = "test.site.ch"
testJob.create(group=testJobGroup)
def createTestJob(self, testSubscription, jobName, *testFiles):
"""
_createTestJob_
Create a test job with two files as input. This will also create the
appropriate workflow, jobgroup and subscription.
"""
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFiles = list(testFiles)
testJob = Job(name=jobName, files=testFiles)
testJob["couch_record"] = "somecouchrecord"
testJob["location"] = "test.site.ch"
testJob.create(group=testJobGroup)
def createDummyJobs(self, nJobs, location=None):
"""
_createDummyJobs_
Create some dummy jobs
"""
if not location:
location = self.sites[0]
nameStr = makeUUID()
testWorkflow = Workflow(spec=nameStr,
owner="tapas",
name=nameStr,
task="basicWorkload/Production",
owner_vogroup='phgroup',
owner_vorole='cmsrole')
testWorkflow.create()
testFileset = Fileset(name=nameStr)
testFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow,
type="Processing",
split_algo="FileBased")
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
jobList = []
for i in range(nJobs):
testJob = Job(name='%s-%i' % (nameStr, i))
testJob['location'] = location
testJob['custom']['location'] = location
testJob['userdn'] = 'tapas'
testJob['owner'] = 'tapas'
testJob['userrole'] = 'cmsrole'
testJob['usergroup'] = 'phgroup'
testJob.create(testJobGroup)
jobList.append(testJob)
return jobList
def testAddChecksumsByLFN(self):
"""
_testAddChecksumsByLFN_
Tests for adding checksums by DAO by LFN
"""
testWorkflow = Workflow(spec="hello", owner="mnorman", name="wf001", task="basicWorkload/Production")
testWorkflow.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testSubscription = Subscription(
fileset=testFileset, workflow=testWorkflow, type="Processing", split_algo="FileBased"
)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileA.addRun(Run(1, *[45]))
testFileA.create()
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileB.addRun(Run(1, *[45]))
testFileB.create()
testJobA = Job()
testJobA.create(group=testJobGroup)
testJobA.associateFiles()
parentAction = self.daofactory(classname="Files.AddChecksumByLFN")
binds = [
{"lfn": testFileA["lfn"], "cktype": "cksum", "cksum": 101},
{"lfn": testFileA["lfn"], "cktype": "adler32", "cksum": 201},
{"lfn": testFileB["lfn"], "cktype": "cksum", "cksum": 101},
]
parentAction.execute(bulkList=binds)
testFileC = File(id=testFileA["id"])
testFileC.loadData()
testFileD = File(id=testFileB["id"])
testFileD.loadData()
self.assertEqual(testFileC["checksums"], {"adler32": "201", "cksum": "101"})
self.assertEqual(testFileD["checksums"], {"cksum": "101"})
return
def testCreateDeleteExists(self):
"""
_testCreateDeleteExists_
Create and then delete a job. Use the job class's exists() method to
determine if the job has been written to the database before it is
created, after it has been created and after it has been deleted.
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Simon",
name="wf001",
task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileA.create()
testFileB.create()
testJob = Job(name="TestJob", files=[testFileA, testFileB])
assert testJob.exists() == False, \
"ERROR: Job exists before it was created"
testJob.create(group=testJobGroup)
assert testJob.exists() >= 0, \
"ERROR: Job does not exist after it was created"
testJob.delete()
assert testJob.exists() == False, \
"ERROR: Job exists after it was delete"
return
def createDummyJobs(self, nJobs, location = None):
"""
_createDummyJobs_
Create some dummy jobs
"""
if not location:
location = self.sites[0]
nameStr = makeUUID()
testWorkflow = Workflow(spec = nameStr, owner = "mnorman",
name = nameStr, task="basicWorkload/Production",
owner_vogroup = 'phgroup', owner_vorole = 'cmsrole')
testWorkflow.create()
testFileset = Fileset(name = nameStr)
testFileset.create()
testSubscription = Subscription(fileset = testFileset,
workflow = testWorkflow,
type = "Processing",
split_algo = "FileBased")
testSubscription.create()
testJobGroup = JobGroup(subscription = testSubscription)
testJobGroup.create()
jobList = []
for i in range(nJobs):
testJob = Job(name = '%s-%i' % (nameStr, i))
testJob['location'] = location
testJob['custom']['location'] = location
testJob['userdn'] = 'mnorman'
testJob['owner'] = 'mnorman'
testJob['userrole'] = 'cmsrole'
testJob['usergroup'] = 'phgroup'
testJob.create(testJobGroup)
jobList.append(testJob)
return jobList
def testCreateDeleteExists(self):
"""
_testCreateDeleteExists_
Create and then delete a job. Use the job class's exists() method to
determine if the job has been written to the database before it is
created, after it has been created and after it has been deleted.
"""
testWorkflow = Workflow(spec = "spec.xml", owner = "Simon",
name = "wf001", task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name = "TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset = testWMBSFileset,
workflow = testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription = testSubscription)
testJobGroup.create()
testFileA = File(lfn = "/this/is/a/lfnA", size = 1024, events = 10)
testFileB = File(lfn = "/this/is/a/lfnB", size = 1024, events = 10)
testFileA.create()
testFileB.create()
testJob = Job(name = "TestJob", files = [testFileA, testFileB])
assert testJob.exists() == False, \
"ERROR: Job exists before it was created"
testJob.create(group = testJobGroup)
assert testJob.exists() >= 0, \
"ERROR: Job does not exist after it was created"
testJob.delete()
assert testJob.exists() == False, \
"ERROR: Job exists after it was delete"
return
def testLoadOutputID(self):
"""
_testLoadOutputID_
Test whether we can load an output ID for a job
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Steve",
name="wf001",
task="Test")
testWorkflow.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow)
testSubscription.create()
testFileA = File(lfn=makeUUID(), locations="test.site.ch")
testFileB = File(lfn=makeUUID(), locations="test.site.ch")
testFileA.create()
testFileB.create()
testFileset.addFile([testFileA, testFileB])
testFileset.commit()
testSubscription.acquireFiles([testFileA, testFileB])
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testJob = Job()
testJob.create(group=testJobGroup)
self.assertEqual(testJob.loadOutputID(), testJobGroup.output.id)
return
def testCreateDeleteExistsNoFiles(self):
"""
_testCreateDeleteExistsNoFiles_
Create and then delete a job but don't add any input files to it.
Use the job class's exists() method to determine if the job has been
written to the database before it is created, after it has been created
and after it has been deleted.
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Simon",
name="wf001",
task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testJob = Job(name="TestJob")
assert testJob.exists() == False, \
"ERROR: Job exists before it was created"
testJob.create(group=testJobGroup)
assert testJob.exists() >= 0, \
"ERROR: Job does not exist after it was created"
testJob.delete()
assert testJob.exists() == False, \
"ERROR: Job exists after it was delete"
return
def testLoadOutputID(self):
"""
_testLoadOutputID_
Test whether we can load an output ID for a job
"""
testWorkflow = Workflow(spec = "spec.xml", owner = "Steve",
name = "wf001", task="Test")
testWorkflow.create()
testFileset = Fileset(name = "TestFileset")
testFileset.create()
testSubscription = Subscription(fileset = testFileset,
workflow = testWorkflow)
testSubscription.create()
testFileA = File(lfn = makeUUID(), locations = "test.site.ch")
testFileB = File(lfn = makeUUID(), locations = "test.site.ch")
testFileA.create()
testFileB.create()
testFileset.addFile([testFileA, testFileB])
testFileset.commit()
testSubscription.acquireFiles([testFileA, testFileB])
testJobGroup = JobGroup(subscription = testSubscription)
testJobGroup.create()
testJob = Job()
testJob.create(group = testJobGroup)
self.assertEqual(testJob.loadOutputID(), testJobGroup.output.id)
return
def testListRunningJobs(self):
"""
_testListRunningJobs_
Test the ListRunningJobs DAO.
"""
testWorkflow = Workflow(spec = makeUUID(), owner = "Steve",
name = makeUUID(), task="Test")
testWorkflow.create()
testFileset = Fileset(name = "TestFileset")
testFileset.create()
testSubscription = Subscription(fileset = testFileset,
workflow = testWorkflow,
type = "Processing")
testSubscription.create()
testJobGroup = JobGroup(subscription = testSubscription)
testJobGroup.create()
testJobA = Job(name = makeUUID(), files = [])
testJobA["couch_record"] = makeUUID()
testJobA.create(group = testJobGroup)
testJobA["state"] = "executing"
testJobB = Job(name = makeUUID(), files = [])
testJobB["couch_record"] = makeUUID()
testJobB.create(group = testJobGroup)
testJobB["state"] = "complete"
testJobC = Job(name = makeUUID(), files = [])
testJobC["couch_record"] = makeUUID()
testJobC.create(group = testJobGroup)
testJobC["state"] = "new"
changeStateAction = self.daoFactory(classname = "Jobs.ChangeState")
changeStateAction.execute(jobs = [testJobA, testJobB, testJobC])
runningJobsAction = self.daoFactory(classname = "Monitoring.ListRunningJobs")
runningJobs = runningJobsAction.execute()
assert len(runningJobs) == 2, \
"Error: Wrong number of running jobs returned."
for runningJob in runningJobs:
if runningJob["job_name"] == testJobA["name"]:
assert runningJob["state"] == testJobA["state"], \
"Error: Running job has wrong state."
assert runningJob["couch_record"] == testJobA["couch_record"], \
"Error: Running job has wrong couch record."
else:
assert runningJob["job_name"] == testJobC["name"], \
"Error: Running job has wrong name."
assert runningJob["state"] == testJobC["state"], \
"Error: Running job has wrong state."
assert runningJob["couch_record"] == testJobC["couch_record"], \
"Error: Running job has wrong couch record."
return
def testCreateTransaction(self):
"""
_testCreateTransaction_
Create a job and save it to the database. Roll back the database
transaction and verify that the job is no longer in the database.
"""
testWorkflow = Workflow(spec="spec.xml", owner="Simon", name="wf001", task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset, workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileA.create()
testFileB.create()
myThread = threading.currentThread()
myThread.transaction.begin()
testJob = Job(name="TestJob", files=[testFileA, testFileB])
assert testJob.exists() == False, "ERROR: Job exists before it was created"
testJob.create(group=testJobGroup)
assert testJob.exists() >= 0, "ERROR: Job does not exist after it was created"
myThread.transaction.rollback()
assert testJob.exists() == False, "ERROR: Job exists after transaction was rolled back."
return
def createTestJob(subscriptionType="Merge"):
"""
_createTestJob_
Create a test job with two files as input. This will also create the
appropriate workflow, jobgroup and subscription.
"""
testWorkflow = Workflow(spec=makeUUID(),
owner="Simon",
name=makeUUID(),
task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow,
type=subscriptionType)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileA.addRun(Run(1, *[45]))
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileB.addRun(Run(1, *[46]))
testFileA.create()
testFileB.create()
testJob = Job(name=makeUUID(), files=[testFileA, testFileB])
testJob["couch_record"] = "somecouchrecord"
testJob["location"] = "test.site.ch"
testJob.create(group=testJobGroup)
testJob.associateFiles()
return testJob
def createJobs(self, nJobs):
"""
Creates a series of jobGroups for submissions
"""
testWorkflow = Workflow(spec="dummy",
owner="mnorman",
name="dummy",
task="basicWorkload/Production")
testWorkflow.create()
# Create Fileset, Subscription, jobGroup
testFileset = Fileset(name="dummy")
testFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow,
type="Processing",
split_algo="FileBased")
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
# Create jobs
for id in range(nJobs):
testJob = Job(name='Job_%i' % (id))
testJob['owner'] = "mnorman"
testJob['location'] = 'Xanadu'
testJob.create(testJobGroup)
testJobGroup.add(testJob)
testFileset.commit()
testJobGroup.commit()
return testJobGroup
def createJobs(self, nJobs):
"""
Creates a series of jobGroups for submissions
"""
testWorkflow = Workflow(spec = "dummy", owner = "mnorman",
name = "dummy", task="basicWorkload/Production")
testWorkflow.create()
# Create Fileset, Subscription, jobGroup
testFileset = Fileset(name = "dummy")
testFileset.create()
testSubscription = Subscription(fileset = testFileset,
workflow = testWorkflow,
type = "Processing",
split_algo = "FileBased")
testSubscription.create()
testJobGroup = JobGroup(subscription = testSubscription)
testJobGroup.create()
# Create jobs
for id in range(nJobs):
testJob = Job(name = 'Job_%i' % (id))
testJob['owner'] = "mnorman"
testJob['location'] = 'Xanadu'
testJob.create(testJobGroup)
testJobGroup.add(testJob)
testFileset.commit()
testJobGroup.commit()
return testJobGroup
def createJobs(self):
"""
_createJobs_
Create test jobs in WMBS and BossAir
"""
testWorkflow = Workflow(spec = makeUUID(), owner = "tapas",
name = makeUUID(), task = "Test")
testWorkflow.create()
testFilesetA = Fileset(name = "TestFilesetA")
testFilesetA.create()
testFilesetB = Fileset(name = "TestFilesetB")
testFilesetB.create()
testFilesetC = Fileset(name = "TestFilesetC")
testFilesetC.create()
testFileA = File(lfn = "testFileA", locations = set(["testSE1", "testSE2"]))
testFileA.create()
testFilesetA.addFile(testFileA)
testFilesetA.commit()
testFilesetB.addFile(testFileA)
testFilesetB.commit()
testFilesetC.addFile(testFileA)
testFilesetC.commit()
testSubscriptionA = Subscription(fileset = testFilesetA,
workflow = testWorkflow,
type = "Processing")
testSubscriptionA.create()
testSubscriptionA.addWhiteBlackList([{"site_name": "testSite1", "valid": True}])
testSubscriptionB = Subscription(fileset = testFilesetB,
workflow = testWorkflow,
type = "Processing")
testSubscriptionB.create()
testSubscriptionB.addWhiteBlackList([{"site_name": "testSite1", "valid": False}])
testSubscriptionC = Subscription(fileset = testFilesetC,
workflow = testWorkflow,
type = "Merge")
testSubscriptionC.create()
testJobGroupA = JobGroup(subscription = testSubscriptionA)
testJobGroupA.create()
testJobGroupB = JobGroup(subscription = testSubscriptionB)
testJobGroupB.create()
testJobGroupC = JobGroup(subscription = testSubscriptionC)
testJobGroupC.create()
# Site1, Has been assigned a location and is complete.
testJobA = Job(name = "testJobA", files = [testFileA])
testJobA["couch_record"] = makeUUID()
testJobA.create(group = testJobGroupA)
testJobA["state"] = "success"
# Site 1, Has been assigned a location and is incomplete.
testJobB = Job(name = "testJobB", files = [testFileA])
testJobB["couch_record"] = makeUUID()
testJobB["cache_dir"] = self.tempDir
testJobB.create(group = testJobGroupA)
testJobB["state"] = "executing"
runJobB = RunJob()
runJobB.buildFromJob(testJobB)
runJobB["status"] = "PEND"
# Does not have a location, white listed to site 1
testJobC = Job(name = "testJobC", files = [testFileA])
testJobC["couch_record"] = makeUUID()
testJobC.create(group = testJobGroupA)
testJobC["state"] = "new"
# Site 2, Has been assigned a location and is complete.
testJobD = Job(name = "testJobD", files = [testFileA])
testJobD["couch_record"] = makeUUID()
testJobD.create(group = testJobGroupB)
testJobD["state"] = "success"
# Site 2, Has been assigned a location and is incomplete.
testJobE = Job(name = "testJobE", files = [testFileA])
testJobE["couch_record"] = makeUUID()
testJobE.create(group = testJobGroupB)
testJobE["state"] = "executing"
runJobE = RunJob()
runJobE.buildFromJob(testJobE)
runJobE["status"] = "RUN"
# Does not have a location, site 1 is blacklisted.
testJobF = Job(name = "testJobF", files = [testFileA])
testJobF["couch_record"] = makeUUID()
testJobF.create(group = testJobGroupB)
testJobF["state"] = "new"
# Site 3, Has been assigned a location and is complete.
testJobG = Job(name = "testJobG", files = [testFileA])
testJobG["couch_record"] = makeUUID()
testJobG.create(group = testJobGroupC)
testJobG["state"] = "cleanout"
# Site 3, Has been assigned a location and is incomplete.
testJobH = Job(name = "testJobH", files = [testFileA])
testJobH["couch_record"] = makeUUID()
testJobH.create(group = testJobGroupC)
testJobH["state"] = "new"
# Site 3, Does not have a location.
testJobI = Job(name = "testJobI", files = [testFileA])
testJobI["couch_record"] = makeUUID()
testJobI.create(group = testJobGroupC)
testJobI["state"] = "new"
# Site 3, Does not have a location and is in cleanout.
testJobJ = Job(name = "testJobJ", files = [testFileA])
testJobJ["couch_record"] = makeUUID()
testJobJ.create(group = testJobGroupC)
testJobJ["state"] = "cleanout"
changeStateAction = self.daoFactory(classname = "Jobs.ChangeState")
changeStateAction.execute(jobs = [testJobA, testJobB, testJobC, testJobD,
testJobE, testJobF, testJobG, testJobH,
testJobI, testJobJ])
self.insertRunJob.execute([runJobB, runJobE])
setLocationAction = self.daoFactory(classname = "Jobs.SetLocation")
setLocationAction.execute(testJobA["id"], "testSite1")
setLocationAction.execute(testJobB["id"], "testSite1")
setLocationAction.execute(testJobD["id"], "testSite1")
setLocationAction.execute(testJobE["id"], "testSite2")
setLocationAction.execute(testJobG["id"], "testSite1")
setLocationAction.execute(testJobH["id"], "testSite1")
return
def createTestJobGroup(self, nJobs = 10, retry_count = 0, workloadPath = 'test'):
"""
Creates a group of several jobs
"""
myThread = threading.currentThread()
myThread.transaction.begin()
testWorkflow = Workflow(spec = workloadPath, owner = "Simon",
name = "wf001", task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name = "TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset = testWMBSFileset,
workflow = testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription = testSubscription)
testJobGroup.create()
testFile0 = File(lfn = "/this/is/a/parent", size = 1024, events = 10)
testFile0.addRun(Run(10, *[12312]))
testFile0.setLocation('malpaquet')
testFileA = File(lfn = "/this/is/a/lfnA", size = 1024, events = 10)
testFileA.addRun(Run(10, *[12312]))
testFileA.setLocation('malpaquet')
testFileB = File(lfn = "/this/is/a/lfnB", size = 1024, events = 10)
testFileB.addRun(Run(10, *[12312]))
testFileB.setLocation('malpaquet')
testFile0.create()
testFileA.create()
testFileB.create()
testFileA.addParent(lfn = "/this/is/a/parent")
testFileB.addParent(lfn = "/this/is/a/parent")
for i in range(0, nJobs):
testJob = Job(name = makeUUID())
testJob['retry_count'] = retry_count
testJob['retry_max'] = 10
testJob['group'] = 'BadGuys'
testJob['user'] = '******'
testJob['taskType'] = 'Merge'
#testJob['fwjr'] = myReport
testJobGroup.add(testJob)
testJob.create(group = testJobGroup)
testJob.addFile(testFileA)
testJob.addFile(testFileB)
testJob.save()
testJobGroup.commit()
testSubscription.acquireFiles(files = [testFileA, testFileB])
testSubscription.save()
myThread.transaction.commit()
return testJobGroup
def injectJobs(self):
"""
_injectJobs_
Inject two workflows into WMBS and save the job objects to disk.
"""
testWorkflowA = Workflow(spec="specA.pkl",
owner="Steve",
name="wf001",
task="TestTaskA")
testWorkflowA.create()
testWorkflowB = Workflow(spec="specB.pkl",
owner="Steve",
name="wf002",
task="TestTaskB")
testWorkflowB.create()
testFileset = Fileset("testFileset")
testFileset.create()
testSubA = Subscription(fileset=testFileset, workflow=testWorkflowA)
testSubA.create()
testSubB = Subscription(fileset=testFileset, workflow=testWorkflowB)
testSubB.create()
testGroupA = JobGroup(subscription=testSubA)
testGroupA.create()
testGroupB = JobGroup(subscription=testSubB)
testGroupB.create()
stateChanger = ChangeState(self.createConfig(),
"jobsubmittercaching_t")
for i in range(10):
newFile = File(lfn="testFile%s" % i,
locations=set(["se.T1_US_FNAL", "se.T1_UK_RAL"]))
newFile.create()
newJobA = Job(name="testJobA-%s" % i, files=[newFile])
newJobA["workflow"] = "wf001"
newJobA["possiblePSN"] = ["T1_US_FNAL"]
newJobA["sandbox"] = "%s/somesandbox" % self.testDir
newJobA["owner"] = "Steve"
jobCacheDir = os.path.join(self.testDir, "jobA-%s" % i)
os.mkdir(jobCacheDir)
newJobA["cache_dir"] = jobCacheDir
newJobA["type"] = "Processing"
newJobA['requestType'] = 'ReReco'
newJobA.create(testGroupA)
jobHandle = open(os.path.join(jobCacheDir, "job.pkl"), "wb")
pickle.dump(newJobA, jobHandle)
jobHandle.close()
stateChanger.propagate([newJobA], "created", "new")
newJobB = Job(name="testJobB-%s" % i, files=[newFile])
newJobB["workflow"] = "wf001"
newJobB["possiblePSN"] = ["T1_UK_RAL"]
newJobB["sandbox"] = "%s/somesandbox" % self.testDir
newJobB["owner"] = "Steve"
jobCacheDir = os.path.join(self.testDir, "jobB-%s" % i)
os.mkdir(jobCacheDir)
newJobB["cache_dir"] = jobCacheDir
newJobB["type"] = "Processing"
newJobB['requestType'] = 'ReReco'
newJobB.create(testGroupB)
jobHandle = open(os.path.join(jobCacheDir, "job.pkl"), "wb")
pickle.dump(newJobB, jobHandle)
jobHandle.close()
stateChanger.propagate([newJobB], "created", "new")
return
def testParallelProcessing(self):
"""
_testParallelProcessing_
Verify that merging works correctly when multiple processing
subscriptions are run over the same input files. The merging algorithm
should ignore processing jobs that feed into different merge
subscriptions.
"""
locationAction = self.daoFactory(classname="Locations.New")
locationAction.execute(siteName="T2_CH_CERN", pnn="T2_CH_CERN")
locationAction.execute(siteName="T1_US_FNAL", pnn="T2_CH_CERN")
mergeFilesetA = Fileset(name="mergeFilesetA")
mergeFilesetB = Fileset(name="mergeFilesetB")
mergeFilesetA.create()
mergeFilesetB.create()
mergeMergedFilesetA = Fileset(name="mergeMergedFilesetA")
mergeMergedFilesetB = Fileset(name="mergeMergedFilesetB")
mergeMergedFilesetA.create()
mergeMergedFilesetB.create()
mergeWorkflow = Workflow(name="mergeWorkflow", spec="bogus",
owner="Steve", task="Test")
mergeWorkflow.create()
mergeSubscriptionA = Subscription(fileset=mergeFilesetA,
workflow=mergeWorkflow,
split_algo="WMBSMergeBySize")
mergeSubscriptionB = Subscription(fileset=mergeFilesetB,
workflow=mergeWorkflow,
split_algo="WMBSMergeBySize")
mergeSubscriptionA.create()
mergeSubscriptionB.create()
inputFileset = Fileset(name="inputFileset")
inputFileset.create()
inputFileA = File(lfn="inputLFNA")
inputFileB = File(lfn="inputLFNB")
inputFileA.create()
inputFileB.create()
procWorkflowA = Workflow(name="procWorkflowA", spec="bunk2",
owner="Steve", task="Test")
procWorkflowA.create()
procWorkflowA.addOutput("output", mergeFilesetA, mergeMergedFilesetA)
procWorkflowB = Workflow(name="procWorkflowB", spec="bunk3",
owner="Steve", task="Test2")
procWorkflowB.create()
procWorkflowB.addOutput("output", mergeFilesetB, mergeMergedFilesetB)
procSubscriptionA = Subscription(fileset=inputFileset,
workflow=procWorkflowA,
split_algo="EventBased")
procSubscriptionA.create()
procSubscriptionB = Subscription(fileset=inputFileset,
workflow=procWorkflowB,
split_algo="EventBased")
procSubscriptionB.create()
jobGroupA = JobGroup(subscription=procSubscriptionA)
jobGroupA.create()
jobGroupB = JobGroup(subscription=procSubscriptionB)
jobGroupB.create()
changeStateDAO = self.daoFactory(classname="Jobs.ChangeState")
testJobA = Job()
testJobA.addFile(inputFileA)
testJobA.create(jobGroupA)
testJobA["state"] = "cleanout"
testJobA["oldstate"] = "new"
testJobA["couch_record"] = "somejive"
testJobA["retry_count"] = 0
testJobA["outcome"] = "success"
testJobA.save()
testJobB = Job()
testJobB.addFile(inputFileB)
testJobB.create(jobGroupA)
testJobB["state"] = "cleanout"
testJobB["oldstate"] = "new"
testJobB["couch_record"] = "somejive"
testJobB["retry_count"] = 0
testJobB["outcome"] = "success"
testJobB.save()
testJobC = Job()
testJobC.addFile(inputFileA)
testJobC.create(jobGroupB)
testJobC["state"] = "cleanout"
testJobC["oldstate"] = "new"
testJobC["couch_record"] = "somejive"
testJobC["retry_count"] = 0
testJobC["outcome"] = "success"
testJobC.save()
testJobD = Job()
testJobD.addFile(inputFileA)
testJobD.create(jobGroupB)
testJobD["state"] = "cleanout"
testJobD["oldstate"] = "new"
testJobD["couch_record"] = "somejive"
testJobD["retry_count"] = 0
testJobD["outcome"] = "failure"
testJobD.save()
testJobE = Job()
testJobE.addFile(inputFileB)
testJobE.create(jobGroupB)
testJobE["state"] = "cleanout"
testJobE["oldstate"] = "new"
testJobE["couch_record"] = "somejive"
testJobE["retry_count"] = 0
testJobE["outcome"] = "success"
testJobE.save()
testJobF = Job()
testJobF.addFile(inputFileB)
testJobF.create(jobGroupB)
testJobF["state"] = "cleanout"
testJobF["oldstate"] = "new"
testJobF["couch_record"] = "somejive"
testJobF["retry_count"] = 0
testJobF["outcome"] = "failure"
testJobF.save()
changeStateDAO.execute([testJobA, testJobB, testJobC, testJobD,
testJobE, testJobF])
fileA = File(lfn="fileA", size=1024, events=1024, first_event=0,
locations={"T2_CH_CERN"})
fileA.addRun(Run(1, *[45]))
fileA.create()
fileA.addParent(inputFileA["lfn"])
fileB = File(lfn="fileB", size=1024, events=1024, first_event=0,
locations={"T2_CH_CERN"})
fileB.addRun(Run(1, *[45]))
fileB.create()
fileB.addParent(inputFileB["lfn"])
jobGroupA.output.addFile(fileA)
jobGroupA.output.addFile(fileB)
jobGroupA.output.commit()
mergeFilesetA.addFile(fileA)
mergeFilesetA.addFile(fileB)
mergeFilesetA.commit()
fileC = File(lfn="fileC", size=1024, events=1024, first_event=0,
locations={"T2_CH_CERN"})
fileC.addRun(Run(1, *[45]))
fileC.create()
fileC.addParent(inputFileA["lfn"])
fileD = File(lfn="fileD", size=1024, events=1024, first_event=0,
locations={"T2_CH_CERN"})
fileD.addRun(Run(1, *[45]))
fileD.create()
fileD.addParent(inputFileB["lfn"])
jobGroupB.output.addFile(fileC)
jobGroupB.output.addFile(fileD)
mergeFilesetB.addFile(fileC)
mergeFilesetB.addFile(fileD)
mergeFilesetB.commit()
splitter = SplitterFactory()
jobFactory = splitter(package="WMCore.WMBS",
subscription=mergeSubscriptionB)
result = jobFactory(min_merge_size=1, max_merge_size=20000,
max_merge_events=7169)
assert len(result) == 0, \
"Error: No merge jobs should have been created."
fileE = File(lfn="fileE", size=1024, events=1024, first_event=0,
locations={"T2_CH_CERN"})
fileE.addRun(Run(1, *[45]))
fileE.create()
fileE.addParent(inputFileA["lfn"])
fileF = File(lfn="fileF", size=1024, events=1024, first_event=0,
locations={"T2_CH_CERN"})
fileF.addRun(Run(1, *[45]))
fileF.create()
fileF.addParent(inputFileB["lfn"])
jobGroupB.output.addFile(fileE)
jobGroupB.output.addFile(fileF)
mergeFilesetB.addFile(fileE)
mergeFilesetB.addFile(fileF)
mergeFilesetB.commit()
testJobD["outcome"] = "success"
testJobD.save()
testJobF["outcome"] = "success"
testJobF.save()
changeStateDAO.execute([testJobD, testJobF])
result = jobFactory(min_merge_size=1, max_merge_size=20000,
max_merge_events=7169)
assert len(result) == 1, \
"Error: One merge job should have been created: %s" % len(result)
return
def testFailJobInput(self):
"""
_testFailJobInput_
Test the Jobs.FailInput DAO and verify that it doesn't affect other
jobs/subscriptions that run over the same files.
"""
testWorkflow = Workflow(spec="spec.xml", owner="Steve",
name="wf001", task="Test")
bogusWorkflow = Workflow(spec="spec1.xml", owner="Steve",
name="wf002", task="Test")
testWorkflow.create()
bogusWorkflow.create()
testFileset = Fileset(name="TestFileset")
bogusFileset = Fileset(name="BogusFileset")
testFileset.create()
bogusFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow)
bogusSubscription = Subscription(fileset=bogusFileset,
workflow=bogusWorkflow)
testSubscription.create()
bogusSubscription.create()
testFileA = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileB = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileC = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileA.create()
testFileB.create()
testFileC.create()
testFileset.addFile([testFileA, testFileB, testFileC])
bogusFileset.addFile([testFileA, testFileB, testFileC])
testFileset.commit()
bogusFileset.commit()
testSubscription.completeFiles([testFileA, testFileB, testFileC])
bogusSubscription.acquireFiles([testFileA, testFileB, testFileC])
testJobGroup = JobGroup(subscription=testSubscription)
bogusJobGroup = JobGroup(subscription=bogusSubscription)
testJobGroup.create()
bogusJobGroup.create()
testJobA = Job(name="TestJobA", files=[testFileA, testFileB, testFileC])
testJobB = Job(name="TestJobB", files=[testFileA, testFileB, testFileC])
bogusJob = Job(name="BogusJob", files=[testFileA, testFileB, testFileC])
testJobA.create(group=testJobGroup)
testJobB.create(group=testJobGroup)
bogusJob.create(group=bogusJobGroup)
testJobA.failInputFiles()
testJobB.failInputFiles()
self.assertEqual(len(testSubscription.filesOfStatus("Available")), 0)
self.assertEqual(len(testSubscription.filesOfStatus("Acquired")), 0)
self.assertEqual(len(testSubscription.filesOfStatus("Failed")), 3)
self.assertEqual(len(testSubscription.filesOfStatus("Completed")), 0)
changeStateAction = self.daoFactory(classname="Jobs.ChangeState")
testJobB["state"] = "cleanout"
changeStateAction.execute([testJobB])
# Try again
testJobA.failInputFiles()
# Should now be failed
self.assertEqual(len(testSubscription.filesOfStatus("Available")), 0)
self.assertEqual(len(testSubscription.filesOfStatus("Acquired")), 0)
self.assertEqual(len(testSubscription.filesOfStatus("Failed")), 3)
self.assertEqual(len(testSubscription.filesOfStatus("Completed")), 0)
# bogus should be unchanged
self.assertEqual(len(bogusSubscription.filesOfStatus("Available")), 0)
self.assertEqual(len(bogusSubscription.filesOfStatus("Acquired")), 3)
self.assertEqual(len(bogusSubscription.filesOfStatus("Failed")), 0)
self.assertEqual(len(bogusSubscription.filesOfStatus("Completed")), 0)
return
def testGetOutputMapDAO(self):
"""
_testGetOutputMapDAO_
Verify the proper behavior of the GetOutputMapDAO for a variety of
different processing chains.
"""
recoOutputFileset = Fileset(name="RECO")
recoOutputFileset.create()
mergedRecoOutputFileset = Fileset(name="MergedRECO")
mergedRecoOutputFileset.create()
alcaOutputFileset = Fileset(name="ALCA")
alcaOutputFileset.create()
mergedAlcaOutputFileset = Fileset(name="MergedALCA")
mergedAlcaOutputFileset.create()
dqmOutputFileset = Fileset(name="DQM")
dqmOutputFileset.create()
mergedDqmOutputFileset = Fileset(name="MergedDQM")
mergedDqmOutputFileset.create()
cleanupFileset = Fileset(name="Cleanup")
cleanupFileset.create()
testWorkflow = Workflow(spec="wf001.xml", owner="Steve",
name="TestWF", task="None")
testWorkflow.create()
testWorkflow.addOutput("output", recoOutputFileset,
mergedRecoOutputFileset)
testWorkflow.addOutput("ALCARECOStreamCombined", alcaOutputFileset,
mergedAlcaOutputFileset)
testWorkflow.addOutput("DQM", dqmOutputFileset,
mergedDqmOutputFileset)
testWorkflow.addOutput("output", cleanupFileset)
testWorkflow.addOutput("ALCARECOStreamCombined", cleanupFileset)
testWorkflow.addOutput("DQM", cleanupFileset)
testRecoMergeWorkflow = Workflow(spec="wf002.xml", owner="Steve",
name="TestRecoMergeWF", task="None")
testRecoMergeWorkflow.create()
testRecoMergeWorkflow.addOutput("anything", mergedRecoOutputFileset,
mergedRecoOutputFileset)
testRecoProcWorkflow = Workflow(spec="wf004.xml", owner="Steve",
name="TestRecoProcWF", task="None")
testRecoProcWorkflow.create()
testAlcaChildWorkflow = Workflow(spec="wf003.xml", owner="Steve",
name="TestAlcaChildWF", task="None")
testAlcaChildWorkflow.create()
inputFile = File(lfn="/path/to/some/lfn", size=600000, events=60000,
locations="cmssrm.fnal.gov")
inputFile.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testFileset.addFile(inputFile)
testFileset.commit()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow,
split_algo="EventBased",
type="Processing")
testMergeRecoSubscription = Subscription(fileset=recoOutputFileset,
workflow=testRecoMergeWorkflow,
split_algo="WMBSMergeBySize",
type="Merge")
testProcRecoSubscription = Subscription(fileset=recoOutputFileset,
workflow=testRecoProcWorkflow,
split_algo="FileBased",
type="Processing")
testChildAlcaSubscription = Subscription(fileset=alcaOutputFileset,
workflow=testAlcaChildWorkflow,
split_algo="FileBased",
type="Processing")
testSubscription.create()
testMergeRecoSubscription.create()
testProcRecoSubscription.create()
testChildAlcaSubscription.create()
testSubscription.acquireFiles()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testJob = Job(name="SplitJobA", files=[inputFile])
testJob.create(group=testJobGroup)
testJob["state"] = "complete"
testJob.save()
outputMapAction = self.daoFactory(classname="Jobs.GetOutputMap")
outputMap = outputMapAction.execute(jobID=testJob["id"])
assert len(outputMap.keys()) == 3, \
"Error: Wrong number of outputs for primary workflow."
goldenMap = {"output": (recoOutputFileset.id,
mergedRecoOutputFileset.id),
"ALCARECOStreamCombined": (alcaOutputFileset.id,
mergedAlcaOutputFileset.id),
"DQM": (dqmOutputFileset.id,
mergedDqmOutputFileset.id)}
for outputID in outputMap.keys():
for outputFilesets in outputMap[outputID]:
if outputFilesets["merged_output_fileset"] is None:
self.assertEqual(outputFilesets["output_fileset"],
cleanupFileset.id,
"Error: Cleanup fileset is wrong.")
continue
self.assertTrue(outputID in goldenMap.keys(),
"Error: Output identifier is missing.")
self.assertEqual(outputFilesets["output_fileset"],
goldenMap[outputID][0],
"Error: Output fileset is wrong.")
self.assertEqual(outputFilesets["merged_output_fileset"],
goldenMap[outputID][1],
"Error: Merged output fileset is wrong.")
del goldenMap[outputID]
self.assertEqual(len(goldenMap.keys()), 0,
"Error: Missing output maps.")
return
def testList(self):
"""
_testList_
Test the functions that list thresholds for creating jobs and submitting
jobs.
"""
myResourceControl = ResourceControl()
myResourceControl.insertSite("testSite1", 10, 20, "testSE1", "testCE1",
"T1_US_FNAL", "LsfPlugin")
myResourceControl.insertSite("testSite2", 20, 40, "testSE2", "testCE2")
myResourceControl.insertThreshold("testSite1", "Processing", 20, 10)
myResourceControl.insertThreshold("testSite1", "Merge", 200, 100)
myResourceControl.insertThreshold("testSite2", "Processing", 50, 25)
myResourceControl.insertThreshold("testSite2", "Merge", 135, 65)
testWorkflow = Workflow(spec=makeUUID(),
owner="Steve",
name=makeUUID(),
task="Test")
testWorkflow.create()
testFilesetA = Fileset(name="TestFilesetA")
testFilesetA.create()
testFilesetB = Fileset(name="TestFilesetB")
testFilesetB.create()
testFilesetC = Fileset(name="TestFilesetC")
testFilesetC.create()
testFileA = File(lfn="testFileA",
locations=set(["testSE1", "testSE2"]))
testFileA.create()
testFilesetA.addFile(testFileA)
testFilesetA.commit()
testFilesetB.addFile(testFileA)
testFilesetB.commit()
testFilesetC.addFile(testFileA)
testFilesetC.commit()
testSubscriptionA = Subscription(fileset=testFilesetA,
workflow=testWorkflow,
type="Processing")
testSubscriptionA.create()
testSubscriptionA.addWhiteBlackList([{
"site_name": "testSite1",
"valid": True
}])
testSubscriptionB = Subscription(fileset=testFilesetB,
workflow=testWorkflow,
type="Processing")
testSubscriptionB.create()
testSubscriptionB.addWhiteBlackList([{
"site_name": "testSite1",
"valid": False
}])
testSubscriptionC = Subscription(fileset=testFilesetC,
workflow=testWorkflow,
type="Merge")
testSubscriptionC.create()
testJobGroupA = JobGroup(subscription=testSubscriptionA)
testJobGroupA.create()
testJobGroupB = JobGroup(subscription=testSubscriptionB)
testJobGroupB.create()
testJobGroupC = JobGroup(subscription=testSubscriptionC)
testJobGroupC.create()
# Site1, Has been assigned a location and is complete.
testJobA = Job(name="testJobA", files=[testFileA])
testJobA["couch_record"] = makeUUID()
testJobA.create(group=testJobGroupA)
testJobA["state"] = "success"
# Site 1, Has been assigned a location and is incomplete.
testJobB = Job(name="testJobB", files=[testFileA])
testJobB["couch_record"] = makeUUID()
testJobB.create(group=testJobGroupA)
testJobB["state"] = "executing"
runJobB = RunJob()
runJobB.buildFromJob(testJobB)
runJobB["status"] = "PEND"
# Does not have a location, white listed to site 1
testJobC = Job(name="testJobC", files=[testFileA])
testJobC["couch_record"] = makeUUID()
testJobC.create(group=testJobGroupA)
testJobC["state"] = "new"
# Site 2, Has been assigned a location and is complete.
testJobD = Job(name="testJobD", files=[testFileA])
testJobD["couch_record"] = makeUUID()
testJobD.create(group=testJobGroupB)
testJobD["state"] = "success"
# Site 2, Has been assigned a location and is incomplete.
testJobE = Job(name="testJobE", files=[testFileA])
testJobE["couch_record"] = makeUUID()
testJobE.create(group=testJobGroupB)
testJobE["state"] = "executing"
runJobE = RunJob()
runJobE.buildFromJob(testJobE)
runJobE["status"] = "RUN"
# Does not have a location, site 1 is blacklisted.
testJobF = Job(name="testJobF", files=[testFileA])
testJobF["couch_record"] = makeUUID()
testJobF.create(group=testJobGroupB)
testJobF["state"] = "new"
# Site 3, Has been assigned a location and is complete.
testJobG = Job(name="testJobG", files=[testFileA])
testJobG["couch_record"] = makeUUID()
testJobG.create(group=testJobGroupC)
testJobG["state"] = "cleanout"
# Site 3, Has been assigned a location and is incomplete.
testJobH = Job(name="testJobH", files=[testFileA])
testJobH["couch_record"] = makeUUID()
testJobH.create(group=testJobGroupC)
testJobH["state"] = "new"
# Site 3, Does not have a location.
testJobI = Job(name="testJobI", files=[testFileA])
testJobI["couch_record"] = makeUUID()
testJobI.create(group=testJobGroupC)
testJobI["state"] = "new"
# Site 3, Does not have a location and is in cleanout.
testJobJ = Job(name="testJobJ", files=[testFileA])
testJobJ["couch_record"] = makeUUID()
testJobJ.create(group=testJobGroupC)
testJobJ["state"] = "cleanout"
changeStateAction = self.daoFactory(classname="Jobs.ChangeState")
changeStateAction.execute(jobs=[
testJobA, testJobB, testJobC, testJobD, testJobE, testJobF,
testJobG, testJobH, testJobI, testJobJ
])
self.insertRunJob.execute([runJobB, runJobE])
setLocationAction = self.daoFactory(classname="Jobs.SetLocation")
setLocationAction.execute(testJobA["id"], "testSite1")
setLocationAction.execute(testJobB["id"], "testSite1")
setLocationAction.execute(testJobD["id"], "testSite1")
setLocationAction.execute(testJobE["id"], "testSite1")
setLocationAction.execute(testJobG["id"], "testSite1")
setLocationAction.execute(testJobH["id"], "testSite1")
createThresholds = myResourceControl.listThresholdsForCreate()
submitThresholds = myResourceControl.listThresholdsForSubmit()
self.assertEqual(len(createThresholds.keys()), 2,
"Error: Wrong number of sites in create thresholds")
self.assertEqual(createThresholds["testSite1"]["total_slots"], 10,
"Error: Wrong number of slots for site 1")
self.assertEqual(createThresholds["testSite2"]["total_slots"], 20,
"Error: Wrong number of slots for site 2")
# We should have two running jobs with locations at site one,
# two running jobs without locations at site two, and one running
# job without a location at site one and two.
self.assertEqual(createThresholds["testSite1"]["pending_jobs"], 4,
"Error: Wrong number of pending jobs for site 1")
# We should have one running job with a location at site 2 and
# another running job without a location.
self.assertEqual(createThresholds["testSite2"]["pending_jobs"], 2,
"Error: Wrong number of pending jobs for site 2")
# We should also have a phedex_name
self.assertEqual(createThresholds["testSite1"]["cms_name"],
"T1_US_FNAL")
self.assertEqual(createThresholds["testSite2"]["cms_name"], None)
mergeThreshold1 = None
mergeThreshold2 = None
procThreshold1 = None
procThreshold2 = None
self.assertEqual(submitThresholds["testSite1"]['cms_name'],
'T1_US_FNAL')
for threshold in submitThresholds["testSite1"]["thresholds"]:
if threshold['task_type'] == "Merge":
mergeThreshold1 = threshold
elif threshold['task_type'] == "Processing":
procThreshold1 = threshold
self.assertEqual(submitThresholds["testSite2"]['cms_name'], None)
for threshold in submitThresholds["testSite2"]["thresholds"]:
if threshold['task_type'] == "Merge":
mergeThreshold2 = threshold
elif threshold['task_type'] == "Processing":
procThreshold2 = threshold
self.assertEqual(
submitThresholds["testSite1"]["total_running_jobs"], 1,
"Error: Wrong number of running jobs for submit thresholds.")
self.assertEqual(
submitThresholds["testSite2"]["total_running_jobs"], 0,
"Error: Wrong number of running jobs for submit thresholds.")
self.assertEqual(
submitThresholds["testSite1"]["total_pending_jobs"], 1,
"Error: Wrong number of pending jobs for submit thresholds.")
self.assertEqual(
submitThresholds["testSite2"]["total_pending_jobs"], 0,
"Error: Wrong number of pending jobs for submit thresholds.")
self.assertEqual(
mergeThreshold1["task_running_jobs"], 0,
"Error: Wrong number of task running jobs for submit thresholds.")
self.assertEqual(
mergeThreshold1["task_pending_jobs"], 0,
"Error: Wrong number of task running jobs for submit thresholds.")
self.assertEqual(
procThreshold1["task_running_jobs"], 1,
"Error: Wrong number of task running jobs for submit thresholds.")
self.assertEqual(
procThreshold1["task_pending_jobs"], 1,
"Error: Wrong number of task running jobs for submit thresholds.")
self.assertEqual(
mergeThreshold2["task_running_jobs"], 0,
"Error: Wrong number of task running jobs for submit thresholds.")
self.assertEqual(
mergeThreshold2["task_pending_jobs"], 0,
"Error: Wrong number of task running jobs for submit thresholds.")
self.assertEqual(
procThreshold2["task_running_jobs"], 0,
"Error: Wrong number of task running jobs for submit thresholds.")
self.assertEqual(
procThreshold2["task_pending_jobs"], 0,
"Error: Wrong number of task running jobs for submit thresholds.")
return
def stuffWMBS(self, injected = True):
"""
_stuffWMBS_
Insert some dummy jobs, jobgroups, filesets, files and subscriptions
into WMBS to test job creation. Three completed job groups each
containing several files are injected. Another incomplete job group is
also injected. Also files are added to the "Mergeable" subscription as
well as to the output fileset for their jobgroups.
"""
locationAction = self.daoFactory(classname = "Locations.New")
locationAction.execute(siteName = "s1", seName = "somese.cern.ch")
changeStateDAO = self.daoFactory(classname = "Jobs.ChangeState")
self.mergeFileset = Fileset(name = "mergeFileset")
self.mergeFileset.create()
self.bogusFileset = Fileset(name = "bogusFileset")
self.bogusFileset.create()
self.mergeMergedFileset = Fileset(name = "mergeMergedFileset")
self.mergeMergedFileset.create()
self.bogusMergedFileset = Fileset(name = "bogusMergedFileset")
self.bogusMergedFileset.create()
mergeWorkflow = Workflow(name = "mergeWorkflow", spec = "bunk2",
owner = "Steve", task="Test")
mergeWorkflow.create()
markWorkflow = self.daoFactory(classname = "Workflow.MarkInjectedWorkflows")
markWorkflow.execute(names = [mergeWorkflow.name], injected = injected)
self.mergeSubscription = Subscription(fileset = self.mergeFileset,
workflow = mergeWorkflow,
split_algo = "WMBSMergeBySize")
self.mergeSubscription.create()
self.bogusSubscription = Subscription(fileset = self.bogusFileset,
workflow = mergeWorkflow,
split_algo = "WMBSMergeBySize")
inputFileset = Fileset(name = "inputFileset")
inputFileset.create()
inputWorkflow = Workflow(name = "inputWorkflow", spec = "input",
owner = "Steve", task = "Test")
inputWorkflow.create()
inputWorkflow.addOutput("output", self.mergeFileset,
self.mergeMergedFileset)
inputWorkflow.addOutput("output2", self.bogusFileset,
self.bogusMergedFileset)
bogusInputWorkflow = Workflow(name = "bogusInputWorkflow", spec = "input",
owner = "Steve", task = "Test")
bogusInputWorkflow.create()
inputSubscription = Subscription(fileset = inputFileset,
workflow = inputWorkflow)
inputSubscription.create()
bogusInputSubscription = Subscription(fileset = inputFileset,
workflow = bogusInputWorkflow)
bogusInputSubscription.create()
parentFile1 = File(lfn = "parentFile1")
parentFile1.create()
parentFile2 = File(lfn = "parentFile2")
parentFile2.create()
parentFile3 = File(lfn = "parentFile3")
parentFile3.create()
parentFile4 = File(lfn = "parentFile4")
parentFile4.create()
self.parentFileSite2 = File(lfn = "parentFileSite2")
self.parentFileSite2.create()
jobGroup1 = JobGroup(subscription = inputSubscription)
jobGroup1.create()
jobGroup2 = JobGroup(subscription = inputSubscription)
jobGroup2.create()
jobGroup3 = JobGroup(subscription = bogusInputSubscription)
jobGroup3.create()
testJob1 = Job()
testJob1.addFile(parentFile1)
testJob1.create(jobGroup1)
testJob1["state"] = "cleanout"
testJob1["oldstate"] = "new"
testJob1["couch_record"] = "somejive"
testJob1["retry_count"] = 0
testJob1["outcome"] = "success"
testJob1.save()
changeStateDAO.execute([testJob1])
testJob1A = Job()
testJob1A.addFile(parentFile1)
testJob1A.create(jobGroup3)
testJob1A["state"] = "cleanout"
testJob1A["oldstate"] = "new"
testJob1A["couch_record"] = "somejive"
testJob1A["retry_count"] = 0
testJob1A["outcome"] = "failure"
testJob1A.save()
changeStateDAO.execute([testJob1A])
testJob2 = Job()
testJob2.addFile(parentFile2)
testJob2.create(jobGroup1)
testJob2["state"] = "cleanout"
testJob2["oldstate"] = "new"
testJob2["couch_record"] = "somejive"
testJob2["retry_count"] = 0
testJob2["outcome"] = "success"
testJob2.save()
changeStateDAO.execute([testJob2])
testJob3 = Job()
testJob3.addFile(parentFile3)
testJob3.create(jobGroup2)
testJob3["state"] = "cleanout"
testJob3["oldstate"] = "new"
testJob3["couch_record"] = "somejive"
testJob3["retry_count"] = 0
testJob3["outcome"] = "success"
testJob3.save()
changeStateDAO.execute([testJob3])
testJob4 = Job()
testJob4.addFile(parentFile4)
testJob4.create(jobGroup2)
testJob4["state"] = "cleanout"
testJob4["oldstate"] = "new"
testJob4["couch_record"] = "somejive"
testJob4["retry_count"] = 0
testJob4["outcome"] = "failure"
testJob4.save()
changeStateDAO.execute([testJob4])
# We'll simulate a failed split by event job that the merger should
# ignore.
parentFile5 = File(lfn = "parentFile5")
parentFile5.create()
testJob5 = Job()
testJob5.addFile(parentFile5)
testJob5.create(jobGroup2)
testJob5["state"] = "cleanout"
testJob5["oldstate"] = "new"
testJob5["couch_record"] = "somejive"
testJob5["retry_count"] = 0
testJob5["outcome"] = "success"
testJob5.save()
changeStateDAO.execute([testJob5])
testJob6 = Job()
testJob6.addFile(parentFile5)
testJob6.create(jobGroup2)
testJob6["state"] = "cleanout"
testJob6["oldstate"] = "new"
testJob6["couch_record"] = "somejive"
testJob6["retry_count"] = 0
testJob6["outcome"] = "failure"
testJob6.save()
changeStateDAO.execute([testJob6])
testJob7 = Job()
testJob7.addFile(self.parentFileSite2)
testJob7.create(jobGroup2)
testJob7["state"] = "cleanout"
testJob7["oldstate"] = "new"
testJob7["couch_record"] = "somejive"
testJob7["retry_count"] = 0
testJob7["outcome"] = "success"
testJob7.save()
changeStateDAO.execute([testJob7])
badFile1 = File(lfn = "badFile1", size = 10241024, events = 10241024,
first_event = 0, locations = set(["somese.cern.ch"]))
badFile1.addRun(Run(1, *[45]))
badFile1.create()
badFile1.addParent(parentFile5["lfn"])
file1 = File(lfn = "file1", size = 1024, events = 1024, first_event = 0,
locations = set(["somese.cern.ch"]))
file1.addRun(Run(1, *[45]))
file1.create()
file1.addParent(parentFile1["lfn"])
file2 = File(lfn = "file2", size = 1024, events = 1024,
first_event = 1024, locations = set(["somese.cern.ch"]))
file2.addRun(Run(1, *[45]))
file2.create()
file2.addParent(parentFile1["lfn"])
file3 = File(lfn = "file3", size = 1024, events = 1024,
first_event = 2048, locations = set(["somese.cern.ch"]))
file3.addRun(Run(1, *[45]))
file3.create()
file3.addParent(parentFile1["lfn"])
file4 = File(lfn = "file4", size = 1024, events = 1024,
first_event = 3072, locations = set(["somese.cern.ch"]))
file4.addRun(Run(1, *[45]))
file4.create()
file4.addParent(parentFile1["lfn"])
fileA = File(lfn = "fileA", size = 1024, events = 1024,
first_event = 0, locations = set(["somese.cern.ch"]))
fileA.addRun(Run(1, *[46]))
fileA.create()
fileA.addParent(parentFile2["lfn"])
fileB = File(lfn = "fileB", size = 1024, events = 1024,
first_event = 1024, locations = set(["somese.cern.ch"]))
fileB.addRun(Run(1, *[46]))
fileB.create()
fileB.addParent(parentFile2["lfn"])
fileC = File(lfn = "fileC", size = 1024, events = 1024,
first_event = 2048, locations = set(["somese.cern.ch"]))
fileC.addRun(Run(1, *[46]))
fileC.create()
fileC.addParent(parentFile2["lfn"])
fileI = File(lfn = "fileI", size = 1024, events = 1024,
first_event = 0, locations = set(["somese.cern.ch"]))
fileI.addRun(Run(2, *[46]))
fileI.create()
fileI.addParent(parentFile3["lfn"])
fileII = File(lfn = "fileII", size = 1024, events = 1024,
first_event = 1024, locations = set(["somese.cern.ch"]))
fileII.addRun(Run(2, *[46]))
fileII.create()
fileII.addParent(parentFile3["lfn"])
fileIII = File(lfn = "fileIII", size = 1024, events = 1024,
first_event = 2048, locations = set(["somese.cern.ch"]))
fileIII.addRun(Run(2, *[46]))
fileIII.create()
fileIII.addParent(parentFile3["lfn"])
fileIV = File(lfn = "fileIV", size = 1024, events = 1024,
first_event = 3072, locations = set(["somese.cern.ch"]))
fileIV.addRun(Run(2, *[46]))
fileIV.create()
fileIV.addParent(parentFile3["lfn"])
fileX = File(lfn = "badFileA", size = 1024, events = 1024,
first_event = 0, locations = set(["somese.cern.ch"]))
fileX.addRun(Run(1, *[47]))
fileX.create()
fileX.addParent(parentFile4["lfn"])
fileY = File(lfn = "badFileB", size = 1024, events = 1024,
first_event = 1024, locations = set(["somese.cern.ch"]))
fileY.addRun(Run(1, *[47]))
fileY.create()
fileY.addParent(parentFile4["lfn"])
fileZ = File(lfn = "badFileC", size = 1024, events = 1024,
first_event = 2048, locations = set(["somese.cern.ch"]))
fileZ.addRun(Run(1, *[47]))
fileZ.create()
fileZ.addParent(parentFile4["lfn"])
jobGroup1.output.addFile(file1)
jobGroup1.output.addFile(file2)
jobGroup1.output.addFile(file3)
jobGroup1.output.addFile(file4)
jobGroup1.output.addFile(fileA)
jobGroup1.output.addFile(fileB)
jobGroup1.output.addFile(fileC)
jobGroup1.output.commit()
jobGroup2.output.addFile(fileI)
jobGroup2.output.addFile(fileII)
jobGroup2.output.addFile(fileIII)
jobGroup2.output.addFile(fileIV)
jobGroup2.output.addFile(fileX)
jobGroup2.output.addFile(fileY)
jobGroup2.output.addFile(fileZ)
jobGroup2.output.addFile(badFile1)
jobGroup2.output.commit()
for file in [file1, file2, file3, file4, fileA, fileB, fileC, fileI,
fileII, fileIII, fileIV, fileX, fileY, fileZ, badFile1]:
self.mergeFileset.addFile(file)
self.bogusFileset.addFile(file)
self.mergeFileset.commit()
self.bogusFileset.commit()
return
def testMask(self):
"""
_testMask_
Test the new mask setup
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Steve",
name="wf001",
task="Test")
testWorkflow.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow)
testSubscription.create()
testFileA = File(lfn=makeUUID(), locations="test.site.ch")
testFileB = File(lfn=makeUUID(), locations="test.site.ch")
testFileA.create()
testFileB.create()
testFileset.addFile([testFileA, testFileB])
testFileset.commit()
testSubscription.acquireFiles([testFileA, testFileB])
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testJob = Job()
testJob['mask'].addRunAndLumis(run=100, lumis=[101, 102])
testJob['mask'].addRunAndLumis(run=200, lumis=[201, 202])
testJob.create(group=testJobGroup)
loadJob = Job(id=testJob.exists())
loadJob.loadData()
runs = loadJob['mask'].getRunAndLumis()
self.assertEqual(len(runs), 2)
self.assertEqual(runs[100], [[101, 102]])
self.assertEqual(runs[200], [[201, 202]])
bigRun = Run(100, *[101, 102, 103, 104])
badRun = Run(300, *[1001, 1002])
result = loadJob['mask'].filterRunLumisByMask([bigRun, badRun])
self.assertEqual(len(result), 1)
alteredRun = result.pop()
self.assertEqual(alteredRun.run, 100)
self.assertEqual(alteredRun.lumis, [101, 102])
run0 = Run(300, *[1001, 1002])
run1 = Run(300, *[1001, 1002])
loadJob['mask'].filterRunLumisByMask([run0, run1])
return
def test_AutoIncrementCheck(self):
"""
_AutoIncrementCheck_
Test and see whether we can find and set the auto_increment values
"""
myThread = threading.currentThread()
if not myThread.dialect.lower() == 'mysql':
return
testWorkflow = Workflow(spec="spec.xml",
owner="Steve",
name="wf001",
task="Test")
testWorkflow.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow)
testSubscription.create()
testFileA = File(lfn=makeUUID(), locations="test.site.ch")
testFileB = File(lfn=makeUUID(), locations="test.site.ch")
testFileA.create()
testFileB.create()
testFileset.addFile([testFileA, testFileB])
testFileset.commit()
testSubscription.acquireFiles([testFileA, testFileB])
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
incrementDAO = self.daoFactory(classname="Jobs.AutoIncrementCheck")
incrementDAO.execute()
testJob = Job()
testJob.create(group=testJobGroup)
self.assertEqual(testJob.exists(), 1)
incrementDAO.execute()
testJob = Job()
testJob.create(group=testJobGroup)
self.assertEqual(testJob.exists(), 2)
incrementDAO.execute(input=10)
testJob = Job()
testJob.create(group=testJobGroup)
self.assertEqual(testJob.exists(), 11)
incrementDAO.execute(input=5)
testJob = Job()
testJob.create(group=testJobGroup)
self.assertEqual(testJob.exists(), 12)
return
def createTestJobGroup(self,
nJobs=10,
retry_count=1,
workloadPath='test',
fwjrPath=None,
workloadName=makeUUID(),
fileModifier=''):
"""
Creates a group of several jobs
"""
myThread = threading.currentThread()
myThread.transaction.begin()
testWorkflow = Workflow(spec=workloadPath,
owner="cmsdataops",
group="cmsdataops",
name=workloadName,
task="/TestWorkload/ReReco")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFile0 = File(lfn="/this/is/a/parent%s" % fileModifier,
size=1024,
events=10)
testFile0.addRun(Run(10, *[12312]))
testFile0.setLocation('T2_CH_CERN')
testFileA = File(lfn="/this/is/a/lfnA%s" % fileModifier,
size=1024,
events=10,
first_event=88,
merged=False)
testFileA.addRun(Run(10, *[12312, 12313]))
testFileA.setLocation('T2_CH_CERN')
testFileB = File(lfn="/this/is/a/lfnB%s" % fileModifier,
size=1024,
events=10,
first_event=88,
merged=False)
testFileB.addRun(Run(10, *[12314, 12315, 12316]))
testFileB.setLocation('T2_CH_CERN')
testFile0.create()
testFileA.create()
testFileB.create()
testFileA.addParent(lfn="/this/is/a/parent%s" % fileModifier)
testFileB.addParent(lfn="/this/is/a/parent%s" % fileModifier)
for i in range(0, nJobs):
testJob = Job(name=makeUUID())
testJob['retry_count'] = retry_count
testJob['retry_max'] = 10
testJob['mask'].addRunAndLumis(run=10, lumis=[12312])
testJob['mask'].addRunAndLumis(run=10, lumis=[12314, 12316])
testJob['cache_dir'] = os.path.join(self.testDir, testJob['name'])
testJob['fwjr_path'] = fwjrPath
os.mkdir(testJob['cache_dir'])
testJobGroup.add(testJob)
testJob.create(group=testJobGroup)
testJob.addFile(testFileA)
testJob.addFile(testFileB)
testJob.save()
testJobGroup.commit()
testSubscription.acquireFiles(files=[testFileA, testFileB])
testSubscription.save()
myThread.transaction.commit()
return testJobGroup
def testParentageByJob(self):
"""
_testParentageByJob_
Tests the DAO that assigns parentage by Job
"""
testWorkflow = Workflow(spec='hello',
owner="mnorman",
name="wf001",
task="basicWorkload/Production")
testWorkflow.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow,
type="Processing",
split_algo="FileBased")
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileParentA = File(lfn="/this/is/a/parent/lfnA",
size=1024,
events=20,
checksums={'cksum': 1})
testFileParentA.addRun(Run(1, *[45]))
testFileParentB = File(lfn="/this/is/a/parent/lfnB",
size=1024,
events=20,
checksums={'cksum': 1})
testFileParentB.addRun(Run(1, *[45]))
testFileParentA.create()
testFileParentB.create()
testFileA = File(lfn="/this/is/a/lfn",
size=1024,
events=10,
checksums={'cksum': 1})
testFileA.addRun(Run(1, *[45]))
testFileA.create()
testJobA = Job()
testJobA.create(group=testJobGroup)
testJobA.addFile(testFileParentA)
testJobA.addFile(testFileParentB)
testJobA.associateFiles()
parentAction = self.daofactory(classname="Files.SetParentageByJob")
parentAction.execute(binds={
'jobid': testJobA.exists(),
'child': testFileA['lfn']
})
testFileB = File(id=testFileA["id"])
testFileB.loadData(parentage=1)
goldenFiles = [testFileParentA, testFileParentB]
for parentFile in testFileB["parents"]:
self.assertEqual(parentFile in goldenFiles, True,
"ERROR: Unknown parent file")
goldenFiles.remove(parentFile)
self.assertEqual(len(goldenFiles), 0,
"ERROR: Some parents are missing")
class JobTestBase(unittest.TestCase):
"""
Base class for tests of WMBS job class (Job_t and JobWorkUnit_t)
"""
def __init__(self, *args, **kwargs):
"""
Define some class instance variables we'll fill later
"""
super(JobTestBase, self).__init__(*args, **kwargs)
self.testFileA = None
self.testFileB = None
self.testWorkflow = None
self.testJob = None
return
def setUp(self):
"""
_setUp_
Setup the database and logging connection. Try to create all of the
WMBS tables.
"""
self.testInit = TestInit(__file__)
self.testInit.setLogging()
self.testInit.setDatabaseConnection()
self.testInit.setSchema(customModules=["WMCore.WMBS"],
useDefault=False)
myThread = threading.currentThread()
self.daoFactory = DAOFactory(package="WMCore.WMBS",
logger=myThread.logger,
dbinterface=myThread.dbi)
locationNew = self.daoFactory(classname="Locations.New")
locationNew.execute(siteName="test.site.ch", pnn="T2_CH_CERN")
locationNew.execute(siteName="test2.site.ch", pnn="T2_CH_CERN")
return
def tearDown(self):
"""
_tearDown_
Drop all the WMBS tables.
"""
self.testInit.clearDatabase()
@staticmethod
def createTestJob(subscriptionType="Merge"):
"""
_createTestJob_
Create a test job with two files as input. This will also create the
appropriate workflow, jobgroup and subscription.
"""
testWorkflow = Workflow(spec=makeUUID(),
owner="Simon",
name=makeUUID(),
task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow,
type=subscriptionType)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileA.addRun(Run(1, *[45]))
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileB.addRun(Run(1, *[46]))
testFileA.create()
testFileB.create()
testJob = Job(name=makeUUID(), files=[testFileA, testFileB])
testJob["couch_record"] = "somecouchrecord"
testJob["location"] = "test.site.ch"
testJob.create(group=testJobGroup)
testJob.associateFiles()
return testJob
def createSingleJobWorkflow(self):
"""
Create a workflow with one jobs and two files and store the results in instance variables
"""
self.testWorkflow = Workflow(spec="spec.xml",
owner="Simon",
name="wf001",
task="Test")
self.testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=self.testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
self.testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
self.testFileA.addRun(Run(1, *[45]))
self.testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
self.testFileB.addRun(Run(1, *[46]))
self.testFileA.create()
self.testFileB.create()
self.testJob = Job(name="TestJob",
files=[self.testFileA, self.testFileB])
self.testJob.create(group=testJobGroup)
self.testJob.associateFiles()
def testParallelProcessing(self):
"""
_testParallelProcessing_
Verify that merging works correctly when multiple processing
subscriptions are run over the same input files. The merging algorithm
should ignore processing jobs that feed into different merge
subscriptions.
"""
locationAction = self.daoFactory(classname = "Locations.New")
locationAction.execute(siteName = "s1", seName = "somese.cern.ch")
mergeFilesetA = Fileset(name = "mergeFilesetA")
mergeFilesetB = Fileset(name = "mergeFilesetB")
mergeFilesetA.create()
mergeFilesetB.create()
mergeMergedFilesetA = Fileset(name = "mergeMergedFilesetA")
mergeMergedFilesetB = Fileset(name = "mergeMergedFilesetB")
mergeMergedFilesetA.create()
mergeMergedFilesetB.create()
mergeWorkflow = Workflow(name = "mergeWorkflow", spec = "bogus",
owner = "Steve", task = "Test")
mergeWorkflow.create()
mergeSubscriptionA = Subscription(fileset = mergeFilesetA,
workflow = mergeWorkflow,
split_algo = "WMBSMergeBySize")
mergeSubscriptionB = Subscription(fileset = mergeFilesetB,
workflow = mergeWorkflow,
split_algo = "WMBSMergeBySize")
mergeSubscriptionA.create()
mergeSubscriptionB.create()
inputFileset = Fileset(name = "inputFileset")
inputFileset.create()
inputFileA = File(lfn = "inputLFNA")
inputFileB = File(lfn = "inputLFNB")
inputFileA.create()
inputFileB.create()
procWorkflowA = Workflow(name = "procWorkflowA", spec = "bunk2",
owner = "Steve", task = "Test")
procWorkflowA.create()
procWorkflowA.addOutput("output", mergeFilesetA, mergeMergedFilesetA)
procWorkflowB = Workflow(name = "procWorkflowB", spec = "bunk3",
owner = "Steve", task = "Test2")
procWorkflowB.create()
procWorkflowB.addOutput("output", mergeFilesetB, mergeMergedFilesetB)
procSubscriptionA = Subscription(fileset = inputFileset,
workflow = procWorkflowA,
split_algo = "EventBased")
procSubscriptionA.create()
procSubscriptionB = Subscription(fileset = inputFileset,
workflow = procWorkflowB,
split_algo = "EventBased")
procSubscriptionB.create()
jobGroupA = JobGroup(subscription = procSubscriptionA)
jobGroupA.create()
jobGroupB = JobGroup(subscription = procSubscriptionB)
jobGroupB.create()
changeStateDAO = self.daoFactory(classname = "Jobs.ChangeState")
testJobA = Job()
testJobA.addFile(inputFileA)
testJobA.create(jobGroupA)
testJobA["state"] = "cleanout"
testJobA["oldstate"] = "new"
testJobA["couch_record"] = "somejive"
testJobA["retry_count"] = 0
testJobA["outcome"] = "success"
testJobA.save()
testJobB = Job()
testJobB.addFile(inputFileB)
testJobB.create(jobGroupA)
testJobB["state"] = "cleanout"
testJobB["oldstate"] = "new"
testJobB["couch_record"] = "somejive"
testJobB["retry_count"] = 0
testJobB["outcome"] = "success"
testJobB.save()
testJobC = Job()
testJobC.addFile(inputFileA)
testJobC.create(jobGroupB)
testJobC["state"] = "cleanout"
testJobC["oldstate"] = "new"
testJobC["couch_record"] = "somejive"
testJobC["retry_count"] = 0
testJobC["outcome"] = "success"
testJobC.save()
testJobD = Job()
testJobD.addFile(inputFileA)
testJobD.create(jobGroupB)
testJobD["state"] = "cleanout"
testJobD["oldstate"] = "new"
testJobD["couch_record"] = "somejive"
testJobD["retry_count"] = 0
testJobD["outcome"] = "failure"
testJobD.save()
testJobE = Job()
testJobE.addFile(inputFileB)
testJobE.create(jobGroupB)
testJobE["state"] = "cleanout"
testJobE["oldstate"] = "new"
testJobE["couch_record"] = "somejive"
testJobE["retry_count"] = 0
testJobE["outcome"] = "success"
testJobE.save()
testJobF = Job()
testJobF.addFile(inputFileB)
testJobF.create(jobGroupB)
testJobF["state"] = "cleanout"
testJobF["oldstate"] = "new"
testJobF["couch_record"] = "somejive"
testJobF["retry_count"] = 0
testJobF["outcome"] = "failure"
testJobF.save()
changeStateDAO.execute([testJobA, testJobB, testJobC, testJobD,
testJobE, testJobF])
fileA = File(lfn = "fileA", size = 1024, events = 1024, first_event = 0,
locations = set(["somese.cern.ch"]))
fileA.addRun(Run(1, *[45]))
fileA.create()
fileA.addParent(inputFileA["lfn"])
fileB = File(lfn = "fileB", size = 1024, events = 1024, first_event = 0,
locations = set(["somese.cern.ch"]))
fileB.addRun(Run(1, *[45]))
fileB.create()
fileB.addParent(inputFileB["lfn"])
jobGroupA.output.addFile(fileA)
jobGroupA.output.addFile(fileB)
jobGroupA.output.commit()
mergeFilesetA.addFile(fileA)
mergeFilesetA.addFile(fileB)
mergeFilesetA.commit()
fileC = File(lfn = "fileC", size = 1024, events = 1024, first_event = 0,
locations = set(["somese.cern.ch"]))
fileC.addRun(Run(1, *[45]))
fileC.create()
fileC.addParent(inputFileA["lfn"])
fileD = File(lfn = "fileD", size = 1024, events = 1024, first_event = 0,
locations = set(["somese.cern.ch"]))
fileD.addRun(Run(1, *[45]))
fileD.create()
fileD.addParent(inputFileB["lfn"])
jobGroupB.output.addFile(fileC)
jobGroupB.output.addFile(fileD)
mergeFilesetB.addFile(fileC)
mergeFilesetB.addFile(fileD)
mergeFilesetB.commit()
splitter = SplitterFactory()
jobFactory = splitter(package = "WMCore.WMBS",
subscription = mergeSubscriptionB)
result = jobFactory(min_merge_size = 1, max_merge_size = 20000,
max_merge_events = 7169)
assert len(result) == 0, \
"Error: No merge jobs should have been created."
fileE = File(lfn = "fileE", size = 1024, events = 1024, first_event = 0,
locations = set(["somese.cern.ch"]))
fileE.addRun(Run(1, *[45]))
fileE.create()
fileE.addParent(inputFileA["lfn"])
fileF = File(lfn = "fileF", size = 1024, events = 1024, first_event = 0,
locations = set(["somese.cern.ch"]))
fileF.addRun(Run(1, *[45]))
fileF.create()
fileF.addParent(inputFileB["lfn"])
jobGroupB.output.addFile(fileE)
jobGroupB.output.addFile(fileF)
mergeFilesetB.addFile(fileE)
mergeFilesetB.addFile(fileF)
mergeFilesetB.commit()
testJobD["outcome"] = "success"
testJobD.save()
testJobF["outcome"] = "success"
testJobF.save()
changeStateDAO.execute([testJobD, testJobF])
result = jobFactory(min_merge_size = 1, max_merge_size = 20000,
max_merge_events = 7169)
assert len(result) == 1, \
"Error: One merge job should have been created: %s" % len(result)
return
def stuffWMBS(self, injected=True):
"""
_stuffWMBS_
Insert some dummy jobs, jobgroups, filesets, files and subscriptions
into WMBS to test job creation. Three completed job groups each
containing several files are injected. Another incomplete job group is
also injected. Also files are added to the "Mergeable" subscription as
well as to the output fileset for their jobgroups.
"""
locationAction = self.daoFactory(classname="Locations.New")
locationAction.execute(siteName="T2_CH_CERN", pnn="T2_CH_CERN")
locationAction.execute(siteName="T1_US_FNAL", pnn="T2_CH_CERN")
changeStateDAO = self.daoFactory(classname="Jobs.ChangeState")
self.mergeFileset = Fileset(name="mergeFileset")
self.mergeFileset.create()
self.bogusFileset = Fileset(name="bogusFileset")
self.bogusFileset.create()
self.mergeMergedFileset = Fileset(name="mergeMergedFileset")
self.mergeMergedFileset.create()
self.bogusMergedFileset = Fileset(name="bogusMergedFileset")
self.bogusMergedFileset.create()
mergeWorkflow = Workflow(name="mergeWorkflow", spec="bunk2",
owner="Steve", task="Test")
mergeWorkflow.create()
markWorkflow = self.daoFactory(classname="Workflow.MarkInjectedWorkflows")
markWorkflow.execute(names=[mergeWorkflow.name], injected=injected)
self.mergeSubscription = Subscription(fileset=self.mergeFileset,
workflow=mergeWorkflow,
split_algo="WMBSMergeBySize")
self.mergeSubscription.create()
self.bogusSubscription = Subscription(fileset=self.bogusFileset,
workflow=mergeWorkflow,
split_algo="WMBSMergeBySize")
inputFileset = Fileset(name="inputFileset")
inputFileset.create()
inputWorkflow = Workflow(name="inputWorkflow", spec="input",
owner="Steve", task="Test")
inputWorkflow.create()
inputWorkflow.addOutput("output", self.mergeFileset,
self.mergeMergedFileset)
inputWorkflow.addOutput("output2", self.bogusFileset,
self.bogusMergedFileset)
bogusInputWorkflow = Workflow(name="bogusInputWorkflow", spec="input",
owner="Steve", task="Test")
bogusInputWorkflow.create()
inputSubscription = Subscription(fileset=inputFileset,
workflow=inputWorkflow)
inputSubscription.create()
bogusInputSubscription = Subscription(fileset=inputFileset,
workflow=bogusInputWorkflow)
bogusInputSubscription.create()
parentFile1 = File(lfn="parentFile1")
parentFile1.create()
parentFile2 = File(lfn="parentFile2")
parentFile2.create()
parentFile3 = File(lfn="parentFile3")
parentFile3.create()
parentFile4 = File(lfn="parentFile4")
parentFile4.create()
self.parentFileSite2 = File(lfn="parentFileSite2")
self.parentFileSite2.create()
jobGroup1 = JobGroup(subscription=inputSubscription)
jobGroup1.create()
jobGroup2 = JobGroup(subscription=inputSubscription)
jobGroup2.create()
jobGroup3 = JobGroup(subscription=bogusInputSubscription)
jobGroup3.create()
testJob1 = Job()
testJob1.addFile(parentFile1)
testJob1.create(jobGroup1)
testJob1["state"] = "cleanout"
testJob1["oldstate"] = "new"
testJob1["couch_record"] = "somejive"
testJob1["retry_count"] = 0
testJob1["outcome"] = "success"
testJob1.save()
changeStateDAO.execute([testJob1])
testJob1A = Job()
testJob1A.addFile(parentFile1)
testJob1A.create(jobGroup3)
testJob1A["state"] = "cleanout"
testJob1A["oldstate"] = "new"
testJob1A["couch_record"] = "somejive"
testJob1A["retry_count"] = 0
testJob1A["outcome"] = "failure"
testJob1A.save()
changeStateDAO.execute([testJob1A])
testJob2 = Job()
testJob2.addFile(parentFile2)
testJob2.create(jobGroup1)
testJob2["state"] = "cleanout"
testJob2["oldstate"] = "new"
testJob2["couch_record"] = "somejive"
testJob2["retry_count"] = 0
testJob2["outcome"] = "success"
testJob2.save()
changeStateDAO.execute([testJob2])
testJob3 = Job()
testJob3.addFile(parentFile3)
testJob3.create(jobGroup2)
testJob3["state"] = "cleanout"
testJob3["oldstate"] = "new"
testJob3["couch_record"] = "somejive"
testJob3["retry_count"] = 0
testJob3["outcome"] = "success"
testJob3.save()
changeStateDAO.execute([testJob3])
testJob4 = Job()
testJob4.addFile(parentFile4)
testJob4.create(jobGroup2)
testJob4["state"] = "cleanout"
testJob4["oldstate"] = "new"
testJob4["couch_record"] = "somejive"
testJob4["retry_count"] = 0
testJob4["outcome"] = "failure"
testJob4.save()
changeStateDAO.execute([testJob4])
# We'll simulate a failed split by event job that the merger should
# ignore.
parentFile5 = File(lfn="parentFile5")
parentFile5.create()
testJob5 = Job()
testJob5.addFile(parentFile5)
testJob5.create(jobGroup2)
testJob5["state"] = "cleanout"
testJob5["oldstate"] = "new"
testJob5["couch_record"] = "somejive"
testJob5["retry_count"] = 0
testJob5["outcome"] = "success"
testJob5.save()
changeStateDAO.execute([testJob5])
testJob6 = Job()
testJob6.addFile(parentFile5)
testJob6.create(jobGroup2)
testJob6["state"] = "cleanout"
testJob6["oldstate"] = "new"
testJob6["couch_record"] = "somejive"
testJob6["retry_count"] = 0
testJob6["outcome"] = "failure"
testJob6.save()
changeStateDAO.execute([testJob6])
testJob7 = Job()
testJob7.addFile(self.parentFileSite2)
testJob7.create(jobGroup2)
testJob7["state"] = "cleanout"
testJob7["oldstate"] = "new"
testJob7["couch_record"] = "somejive"
testJob7["retry_count"] = 0
testJob7["outcome"] = "success"
testJob7.save()
changeStateDAO.execute([testJob7])
badFile1 = File(lfn="badFile1", size=10241024, events=10241024,
first_event=0, locations={"T2_CH_CERN"})
badFile1.addRun(Run(1, *[45]))
badFile1.create()
badFile1.addParent(parentFile5["lfn"])
file1 = File(lfn="file1", size=1024, events=1024, first_event=0,
locations={"T2_CH_CERN"})
file1.addRun(Run(1, *[45]))
file1.create()
file1.addParent(parentFile1["lfn"])
file2 = File(lfn="file2", size=1024, events=1024,
first_event=1024, locations={"T2_CH_CERN"})
file2.addRun(Run(1, *[45]))
file2.create()
file2.addParent(parentFile1["lfn"])
file3 = File(lfn="file3", size=1024, events=1024,
first_event=2048, locations={"T2_CH_CERN"})
file3.addRun(Run(1, *[45]))
file3.create()
file3.addParent(parentFile1["lfn"])
file4 = File(lfn="file4", size=1024, events=1024,
first_event=3072, locations={"T2_CH_CERN"})
file4.addRun(Run(1, *[45]))
file4.create()
file4.addParent(parentFile1["lfn"])
fileA = File(lfn="fileA", size=1024, events=1024,
first_event=0, locations={"T2_CH_CERN"})
fileA.addRun(Run(1, *[46]))
fileA.create()
fileA.addParent(parentFile2["lfn"])
fileB = File(lfn="fileB", size=1024, events=1024,
first_event=1024, locations={"T2_CH_CERN"})
fileB.addRun(Run(1, *[46]))
fileB.create()
fileB.addParent(parentFile2["lfn"])
fileC = File(lfn="fileC", size=1024, events=1024,
first_event=2048, locations={"T2_CH_CERN"})
fileC.addRun(Run(1, *[46]))
fileC.create()
fileC.addParent(parentFile2["lfn"])
fileI = File(lfn="fileI", size=1024, events=1024,
first_event=0, locations={"T2_CH_CERN"})
fileI.addRun(Run(2, *[46]))
fileI.create()
fileI.addParent(parentFile3["lfn"])
fileII = File(lfn="fileII", size=1024, events=1024,
first_event=1024, locations={"T2_CH_CERN"})
fileII.addRun(Run(2, *[46]))
fileII.create()
fileII.addParent(parentFile3["lfn"])
fileIII = File(lfn="fileIII", size=1024, events=1024,
first_event=2048, locations={"T2_CH_CERN"})
fileIII.addRun(Run(2, *[46]))
fileIII.create()
fileIII.addParent(parentFile3["lfn"])
fileIV = File(lfn="fileIV", size=1024, events=1024,
first_event=3072, locations={"T2_CH_CERN"})
fileIV.addRun(Run(2, *[46]))
fileIV.create()
fileIV.addParent(parentFile3["lfn"])
fileX = File(lfn="badFileA", size=1024, events=1024,
first_event=0, locations={"T2_CH_CERN"})
fileX.addRun(Run(1, *[47]))
fileX.create()
fileX.addParent(parentFile4["lfn"])
fileY = File(lfn="badFileB", size=1024, events=1024,
first_event=1024, locations={"T2_CH_CERN"})
fileY.addRun(Run(1, *[47]))
fileY.create()
fileY.addParent(parentFile4["lfn"])
fileZ = File(lfn="badFileC", size=1024, events=1024,
first_event=2048, locations={"T2_CH_CERN"})
fileZ.addRun(Run(1, *[47]))
fileZ.create()
fileZ.addParent(parentFile4["lfn"])
jobGroup1.output.addFile(file1)
jobGroup1.output.addFile(file2)
jobGroup1.output.addFile(file3)
jobGroup1.output.addFile(file4)
jobGroup1.output.addFile(fileA)
jobGroup1.output.addFile(fileB)
jobGroup1.output.addFile(fileC)
jobGroup1.output.commit()
jobGroup2.output.addFile(fileI)
jobGroup2.output.addFile(fileII)
jobGroup2.output.addFile(fileIII)
jobGroup2.output.addFile(fileIV)
jobGroup2.output.addFile(fileX)
jobGroup2.output.addFile(fileY)
jobGroup2.output.addFile(fileZ)
jobGroup2.output.addFile(badFile1)
jobGroup2.output.commit()
for fileObj in [file1, file2, file3, file4, fileA, fileB, fileC, fileI,
fileII, fileIII, fileIV, fileX, fileY, fileZ, badFile1]:
self.mergeFileset.addFile(fileObj)
self.bogusFileset.addFile(fileObj)
self.mergeFileset.commit()
self.bogusFileset.commit()
return
def testListRunningJobs(self):
"""
_testListRunningJobs_
Test the ListRunningJobs DAO.
"""
testWorkflow = Workflow(spec=makeUUID(),
owner="Steve",
name=makeUUID(),
task="Test")
testWorkflow.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow,
type="Processing")
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testJobA = Job(name=makeUUID(), files=[])
testJobA["couch_record"] = makeUUID()
testJobA.create(group=testJobGroup)
testJobA["state"] = "executing"
testJobB = Job(name=makeUUID(), files=[])
testJobB["couch_record"] = makeUUID()
testJobB.create(group=testJobGroup)
testJobB["state"] = "complete"
testJobC = Job(name=makeUUID(), files=[])
testJobC["couch_record"] = makeUUID()
testJobC.create(group=testJobGroup)
testJobC["state"] = "new"
changeStateAction = self.daoFactory(classname="Jobs.ChangeState")
changeStateAction.execute(jobs=[testJobA, testJobB, testJobC])
runningJobsAction = self.daoFactory(
classname="Monitoring.ListRunningJobs")
runningJobs = runningJobsAction.execute()
assert len(runningJobs) == 2, \
"Error: Wrong number of running jobs returned."
for runningJob in runningJobs:
if runningJob["job_name"] == testJobA["name"]:
assert runningJob["state"] == testJobA["state"], \
"Error: Running job has wrong state."
assert runningJob["couch_record"] == testJobA["couch_record"], \
"Error: Running job has wrong couch record."
else:
assert runningJob["job_name"] == testJobC["name"], \
"Error: Running job has wrong name."
assert runningJob["state"] == testJobC["state"], \
"Error: Running job has wrong state."
assert runningJob["couch_record"] == testJobC["couch_record"], \
"Error: Running job has wrong couch record."
return
def testGetOutputParentLFNs(self):
"""
_testGetOutputParentLFNs_
Verify that the getOutputDBSParentLFNs() method returns the correct
parent LFNs.
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Simon",
name="wf001",
task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA",
size=1024,
events=10,
merged=True)
testFileB = File(lfn="/this/is/a/lfnB",
size=1024,
events=10,
merged=True)
testFileC = File(lfn="/this/is/a/lfnC",
size=1024,
events=10,
merged=False)
testFileD = File(lfn="/this/is/a/lfnD",
size=1024,
events=10,
merged=False)
testFileE = File(lfn="/this/is/a/lfnE",
size=1024,
events=10,
merged=True)
testFileF = File(lfn="/this/is/a/lfnF",
size=1024,
events=10,
merged=True)
testFileA.create()
testFileB.create()
testFileC.create()
testFileD.create()
testFileE.create()
testFileF.create()
testFileE.addChild(testFileC["lfn"])
testFileF.addChild(testFileD["lfn"])
testJobA = Job(name="TestJob", files=[testFileA, testFileB])
testJobA["couch_record"] = "somecouchrecord"
testJobA["location"] = "test.site.ch"
testJobA.create(group=testJobGroup)
testJobA.associateFiles()
testJobB = Job(name="TestJobB", files=[testFileC, testFileD])
testJobB["couch_record"] = "somecouchrecord"
testJobB["location"] = "test.site.ch"
testJobB.create(group=testJobGroup)
testJobB.associateFiles()
goldenLFNs = ["/this/is/a/lfnA", "/this/is/a/lfnB"]
parentLFNs = testJobA.getOutputDBSParentLFNs()
for parentLFN in parentLFNs:
assert parentLFN in goldenLFNs, \
"ERROR: Unknown lfn: %s" % parentLFN
goldenLFNs.remove(parentLFN)
assert len(goldenLFNs) == 0, \
"ERROR: LFNs are missing: %s" % goldenLFNs
goldenLFNs = ["/this/is/a/lfnE", "/this/is/a/lfnF"]
parentLFNs = testJobB.getOutputDBSParentLFNs()
for parentLFN in parentLFNs:
assert parentLFN in goldenLFNs, \
"ERROR: Unknown lfn: %s" % parentLFN
goldenLFNs.remove(parentLFN)
assert len(goldenLFNs) == 0, \
"ERROR: LFNs are missing..."
return
def testGetOutputParentLFNs(self):
"""
_testGetOutputParentLFNs_
Verify that the getOutputDBSParentLFNs() method returns the correct
parent LFNs.
"""
testWorkflow = Workflow(spec="spec.xml", owner="Simon",
name="wf001", task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10,
merged=True)
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10,
merged=True)
testFileC = File(lfn="/this/is/a/lfnC", size=1024, events=10,
merged=False)
testFileD = File(lfn="/this/is/a/lfnD", size=1024, events=10,
merged=False)
testFileE = File(lfn="/this/is/a/lfnE", size=1024, events=10,
merged=True)
testFileF = File(lfn="/this/is/a/lfnF", size=1024, events=10,
merged=True)
testFileA.create()
testFileB.create()
testFileC.create()
testFileD.create()
testFileE.create()
testFileF.create()
testFileE.addChild(testFileC["lfn"])
testFileF.addChild(testFileD["lfn"])
testJobA = Job(name="TestJob", files=[testFileA, testFileB])
testJobA["couch_record"] = "somecouchrecord"
testJobA["location"] = "test.site.ch"
testJobA.create(group=testJobGroup)
testJobA.associateFiles()
testJobB = Job(name="TestJobB", files=[testFileC, testFileD])
testJobB["couch_record"] = "somecouchrecord"
testJobB["location"] = "test.site.ch"
testJobB.create(group=testJobGroup)
testJobB.associateFiles()
goldenLFNs = ["/this/is/a/lfnA", "/this/is/a/lfnB"]
parentLFNs = testJobA.getOutputDBSParentLFNs()
for parentLFN in parentLFNs:
assert parentLFN in goldenLFNs, \
"ERROR: Unknown lfn: %s" % parentLFN
goldenLFNs.remove(parentLFN)
assert not goldenLFNs, "ERROR: LFNs are missing: %s" % goldenLFNs
goldenLFNs = ["/this/is/a/lfnE", "/this/is/a/lfnF"]
parentLFNs = testJobB.getOutputDBSParentLFNs()
for parentLFN in parentLFNs:
assert parentLFN in goldenLFNs, \
"ERROR: Unknown lfn: %s" % parentLFN
goldenLFNs.remove(parentLFN)
assert not goldenLFNs, "ERROR: LFNs are missing..."
return
def testFailJobInput(self):
"""
_testFailJobInput_
Test the Jobs.FailInput DAO and verify that it doesn't affect other
jobs/subscriptions that run over the same files.
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Steve",
name="wf001",
task="Test")
bogusWorkflow = Workflow(spec="spec1.xml",
owner="Steve",
name="wf002",
task="Test")
testWorkflow.create()
bogusWorkflow.create()
testFileset = Fileset(name="TestFileset")
bogusFileset = Fileset(name="BogusFileset")
testFileset.create()
bogusFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow)
bogusSubscription = Subscription(fileset=bogusFileset,
workflow=bogusWorkflow)
testSubscription.create()
bogusSubscription.create()
testFileA = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileB = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileC = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileA.create()
testFileB.create()
testFileC.create()
testFileset.addFile([testFileA, testFileB, testFileC])
bogusFileset.addFile([testFileA, testFileB, testFileC])
testFileset.commit()
bogusFileset.commit()
testSubscription.completeFiles([testFileA, testFileB, testFileC])
bogusSubscription.acquireFiles([testFileA, testFileB, testFileC])
testJobGroup = JobGroup(subscription=testSubscription)
bogusJobGroup = JobGroup(subscription=bogusSubscription)
testJobGroup.create()
bogusJobGroup.create()
testJobA = Job(name="TestJobA",
files=[testFileA, testFileB, testFileC])
testJobB = Job(name="TestJobB",
files=[testFileA, testFileB, testFileC])
bogusJob = Job(name="BogusJob",
files=[testFileA, testFileB, testFileC])
testJobA.create(group=testJobGroup)
testJobB.create(group=testJobGroup)
bogusJob.create(group=bogusJobGroup)
testJobA.failInputFiles()
testJobB.failInputFiles()
self.assertEqual(len(testSubscription.filesOfStatus("Available")), 0)
self.assertEqual(len(testSubscription.filesOfStatus("Acquired")), 0)
self.assertEqual(len(testSubscription.filesOfStatus("Failed")), 3)
self.assertEqual(len(testSubscription.filesOfStatus("Completed")), 0)
changeStateAction = self.daoFactory(classname="Jobs.ChangeState")
testJobB["state"] = "cleanout"
changeStateAction.execute([testJobB])
# Try again
testJobA.failInputFiles()
# Should now be failed
self.assertEqual(len(testSubscription.filesOfStatus("Available")), 0)
self.assertEqual(len(testSubscription.filesOfStatus("Acquired")), 0)
self.assertEqual(len(testSubscription.filesOfStatus("Failed")), 3)
self.assertEqual(len(testSubscription.filesOfStatus("Completed")), 0)
# bogus should be unchanged
self.assertEqual(len(bogusSubscription.filesOfStatus("Available")), 0)
self.assertEqual(len(bogusSubscription.filesOfStatus("Acquired")), 3)
self.assertEqual(len(bogusSubscription.filesOfStatus("Failed")), 0)
self.assertEqual(len(bogusSubscription.filesOfStatus("Completed")), 0)
return
def testCompleteJobInput(self):
"""
_testCompleteJobInput_
Verify the correct output of the CompleteInput DAO. This should mark
the input for a job as complete once all the jobs that run over a
particular file have complete successfully.
"""
testWorkflow = Workflow(spec="spec.xml", owner="Steve",
name="wf001", task="Test")
bogusWorkflow = Workflow(spec="spec1.xml", owner="Steve",
name="wf002", task="Test")
testWorkflow.create()
bogusWorkflow.create()
testFileset = Fileset(name="TestFileset")
bogusFileset = Fileset(name="BogusFileset")
testFileset.create()
bogusFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow)
bogusSubscription = Subscription(fileset=bogusFileset,
workflow=bogusWorkflow)
testSubscription.create()
bogusSubscription.create()
testFileA = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileB = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileC = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileA.create()
testFileB.create()
testFileC.create()
testFileset.addFile([testFileA, testFileB, testFileC])
bogusFileset.addFile([testFileA, testFileB, testFileC])
testFileset.commit()
bogusFileset.commit()
testSubscription.acquireFiles([testFileA, testFileB, testFileC])
bogusSubscription.acquireFiles([testFileA, testFileB, testFileC])
testJobGroup = JobGroup(subscription=testSubscription)
bogusJobGroup = JobGroup(subscription=bogusSubscription)
testJobGroup.create()
bogusJobGroup.create()
testJobA = Job(name="TestJobA", files=[testFileA])
testJobB = Job(name="TestJobB", files=[testFileA, testFileB])
testJobC = Job(name="TestJobC", files=[testFileC])
bogusJob = Job(name="BogusJob", files=[testFileA, testFileB, testFileC])
testJobA.create(group=testJobGroup)
testJobB.create(group=testJobGroup)
testJobC.create(group=testJobGroup)
bogusJob.create(group=bogusJobGroup)
testJobA["outcome"] = "success"
testJobB["outcome"] = "failure"
testJobC["outcome"] = "success"
testJobA.save()
testJobB.save()
testJobC.save()
testJobA.completeInputFiles()
compFiles = len(testSubscription.filesOfStatus("Completed"))
assert compFiles == 0, \
"Error: test sub has wrong number of complete files: %s" % compFiles
testJobB["outcome"] = "success"
testJobB.save()
testJobB.completeInputFiles(skipFiles=[testFileB["lfn"]])
availFiles = len(testSubscription.filesOfStatus("Available"))
assert availFiles == 0, \
"Error: test sub has wrong number of available files: %s" % availFiles
acqFiles = len(testSubscription.filesOfStatus("Acquired"))
assert acqFiles == 1, \
"Error: test sub has wrong number of acquired files: %s" % acqFiles
compFiles = len(testSubscription.filesOfStatus("Completed"))
assert compFiles == 1, \
"Error: test sub has wrong number of complete files: %s" % compFiles
failFiles = len(testSubscription.filesOfStatus("Failed"))
assert failFiles == 1, \
"Error: test sub has wrong number of failed files: %s" % failFiles
availFiles = len(bogusSubscription.filesOfStatus("Available"))
assert availFiles == 0, \
"Error: test sub has wrong number of available files: %s" % availFiles
acqFiles = len(bogusSubscription.filesOfStatus("Acquired"))
assert acqFiles == 3, \
"Error: test sub has wrong number of acquired files: %s" % acqFiles
compFiles = len(bogusSubscription.filesOfStatus("Completed"))
assert compFiles == 0, \
"Error: test sub has wrong number of complete files: %s" % compFiles
failFiles = len(bogusSubscription.filesOfStatus("Failed"))
assert failFiles == 0, \
"Error: test sub has wrong number of failed files: %s" % failFiles
return
def testCompleteJobInput(self):
"""
_testCompleteJobInput_
Verify the correct output of the CompleteInput DAO. This should mark
the input for a job as complete once all the jobs that run over a
particular file have complete successfully.
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Steve",
name="wf001",
task="Test")
bogusWorkflow = Workflow(spec="spec1.xml",
owner="Steve",
name="wf002",
task="Test")
testWorkflow.create()
bogusWorkflow.create()
testFileset = Fileset(name="TestFileset")
bogusFileset = Fileset(name="BogusFileset")
testFileset.create()
bogusFileset.create()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow)
bogusSubscription = Subscription(fileset=bogusFileset,
workflow=bogusWorkflow)
testSubscription.create()
bogusSubscription.create()
testFileA = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileB = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileC = File(lfn=makeUUID(), locations="T2_CH_CERN")
testFileA.create()
testFileB.create()
testFileC.create()
testFileset.addFile([testFileA, testFileB, testFileC])
bogusFileset.addFile([testFileA, testFileB, testFileC])
testFileset.commit()
bogusFileset.commit()
testSubscription.acquireFiles([testFileA, testFileB, testFileC])
bogusSubscription.acquireFiles([testFileA, testFileB, testFileC])
testJobGroup = JobGroup(subscription=testSubscription)
bogusJobGroup = JobGroup(subscription=bogusSubscription)
testJobGroup.create()
bogusJobGroup.create()
testJobA = Job(name="TestJobA", files=[testFileA])
testJobB = Job(name="TestJobB", files=[testFileA, testFileB])
testJobC = Job(name="TestJobC", files=[testFileC])
bogusJob = Job(name="BogusJob",
files=[testFileA, testFileB, testFileC])
testJobA.create(group=testJobGroup)
testJobB.create(group=testJobGroup)
testJobC.create(group=testJobGroup)
bogusJob.create(group=bogusJobGroup)
testJobA["outcome"] = "success"
testJobB["outcome"] = "failure"
testJobC["outcome"] = "success"
testJobA.save()
testJobB.save()
testJobC.save()
testJobA.completeInputFiles()
compFiles = len(testSubscription.filesOfStatus("Completed"))
assert compFiles == 0, \
"Error: test sub has wrong number of complete files: %s" % compFiles
testJobB["outcome"] = "success"
testJobB.save()
testJobB.completeInputFiles(skipFiles=[testFileB["lfn"]])
availFiles = len(testSubscription.filesOfStatus("Available"))
assert availFiles == 0, \
"Error: test sub has wrong number of available files: %s" % availFiles
acqFiles = len(testSubscription.filesOfStatus("Acquired"))
assert acqFiles == 1, \
"Error: test sub has wrong number of acquired files: %s" % acqFiles
compFiles = len(testSubscription.filesOfStatus("Completed"))
assert compFiles == 1, \
"Error: test sub has wrong number of complete files: %s" % compFiles
failFiles = len(testSubscription.filesOfStatus("Failed"))
assert failFiles == 1, \
"Error: test sub has wrong number of failed files: %s" % failFiles
availFiles = len(bogusSubscription.filesOfStatus("Available"))
assert availFiles == 0, \
"Error: test sub has wrong number of available files: %s" % availFiles
acqFiles = len(bogusSubscription.filesOfStatus("Acquired"))
assert acqFiles == 3, \
"Error: test sub has wrong number of acquired files: %s" % acqFiles
compFiles = len(bogusSubscription.filesOfStatus("Completed"))
assert compFiles == 0, \
"Error: test sub has wrong number of complete files: %s" % compFiles
failFiles = len(bogusSubscription.filesOfStatus("Failed"))
assert failFiles == 0, \
"Error: test sub has wrong number of failed files: %s" % failFiles
return
class WMBSHelperTest(unittest.TestCase):
def setUp(self):
"""
_setUp_
"""
self.testInit = TestInitCouchApp(__file__)
self.testInit.setLogging()
self.testInit.setDatabaseConnection()
self.testInit.setupCouch("wmbshelper_t/jobs", "JobDump")
self.testInit.setupCouch("wmbshelper_t/fwjrs", "FWJRDump")
self.testInit.setupCouch("config_test", "GroupUser", "ConfigCache")
os.environ["COUCHDB"] = "wmbshelper_t"
self.testInit.setSchema(customModules=[
"WMCore.WMBS", "WMComponent.DBS3Buffer", "WMCore.BossAir",
"WMCore.ResourceControl"
],
useDefault=False)
self.workDir = self.testInit.generateWorkDir()
self.wmspec = self.createWMSpec()
self.topLevelTask = getFirstTask(self.wmspec)
self.inputDataset = self.topLevelTask.inputDataset()
self.dataset = self.topLevelTask.getInputDatasetPath()
self.dbs = MockDBSReader(self.inputDataset.dbsurl)
self.daoFactory = DAOFactory(package="WMCore.WMBS",
logger=threading.currentThread().logger,
dbinterface=threading.currentThread().dbi)
return
def tearDown(self):
"""
_tearDown_
Clear out the database.
"""
self.testInit.clearDatabase()
self.testInit.tearDownCouch()
self.testInit.delWorkDir()
return
def setupForKillTest(self, baAPI=None):
"""
_setupForKillTest_
Inject a workflow into WMBS that has a processing task, a merge task and
a cleanup task. Inject files into the various tasks at various
processing states (acquired, complete, available...). Also create jobs
for each subscription in various states.
"""
myThread = threading.currentThread()
daoFactory = DAOFactory(package="WMCore.WMBS",
logger=myThread.logger,
dbinterface=myThread.dbi)
locationAction = daoFactory(classname="Locations.New")
changeStateAction = daoFactory(classname="Jobs.ChangeState")
resourceControl = ResourceControl()
resourceControl.insertSite(siteName='site1',
seName='goodse.cern.ch',
ceName='site1',
plugin="TestPlugin")
resourceControl.insertThreshold(siteName = 'site1', taskType = 'Processing', \
maxSlots = 10000, pendingSlots = 10000)
userDN = 'someDN'
userAction = daoFactory(classname="Users.New")
userAction.execute(dn=userDN,
group_name='DEFAULT',
role_name='DEFAULT')
inputFileset = Fileset("input")
inputFileset.create()
inputFileA = File("lfnA", locations="goodse.cern.ch")
inputFileB = File("lfnB", locations="goodse.cern.ch")
inputFileC = File("lfnC", locations="goodse.cern.ch")
inputFileA.create()
inputFileB.create()
inputFileC.create()
inputFileset.addFile(inputFileA)
inputFileset.addFile(inputFileB)
inputFileset.addFile(inputFileC)
inputFileset.commit()
unmergedOutputFileset = Fileset("unmerged")
unmergedOutputFileset.create()
unmergedFileA = File("ulfnA", locations="goodse.cern.ch")
unmergedFileB = File("ulfnB", locations="goodse.cern.ch")
unmergedFileC = File("ulfnC", locations="goodse.cern.ch")
unmergedFileA.create()
unmergedFileB.create()
unmergedFileC.create()
unmergedOutputFileset.addFile(unmergedFileA)
unmergedOutputFileset.addFile(unmergedFileB)
unmergedOutputFileset.addFile(unmergedFileC)
unmergedOutputFileset.commit()
mainProcWorkflow = Workflow(spec="spec1",
owner="Steve",
name="Main",
task="Proc")
mainProcWorkflow.create()
mainProcMergeWorkflow = Workflow(spec="spec1",
owner="Steve",
name="Main",
task="ProcMerge")
mainProcMergeWorkflow.create()
mainCleanupWorkflow = Workflow(spec="spec1",
owner="Steve",
name="Main",
task="Cleanup")
mainCleanupWorkflow.create()
self.mainProcSub = Subscription(fileset=inputFileset,
workflow=mainProcWorkflow,
type="Processing")
self.mainProcSub.create()
self.mainProcSub.acquireFiles(inputFileA)
self.mainProcSub.completeFiles(inputFileB)
procJobGroup = JobGroup(subscription=self.mainProcSub)
procJobGroup.create()
self.procJobA = Job(name="ProcJobA")
self.procJobA["state"] = "new"
self.procJobA["location"] = "site1"
self.procJobB = Job(name="ProcJobB")
self.procJobB["state"] = "executing"
self.procJobB["location"] = "site1"
self.procJobC = Job(name="ProcJobC")
self.procJobC["state"] = "complete"
self.procJobC["location"] = "site1"
self.procJobA.create(procJobGroup)
self.procJobB.create(procJobGroup)
self.procJobC.create(procJobGroup)
self.mainMergeSub = Subscription(fileset=unmergedOutputFileset,
workflow=mainProcMergeWorkflow,
type="Merge")
self.mainMergeSub.create()
self.mainMergeSub.acquireFiles(unmergedFileA)
self.mainMergeSub.failFiles(unmergedFileB)
mergeJobGroup = JobGroup(subscription=self.mainMergeSub)
mergeJobGroup.create()
self.mergeJobA = Job(name="MergeJobA")
self.mergeJobA["state"] = "exhausted"
self.mergeJobA["location"] = "site1"
self.mergeJobB = Job(name="MergeJobB")
self.mergeJobB["state"] = "cleanout"
self.mergeJobB["location"] = "site1"
self.mergeJobC = Job(name="MergeJobC")
self.mergeJobC["state"] = "new"
self.mergeJobC["location"] = "site1"
self.mergeJobA.create(mergeJobGroup)
self.mergeJobB.create(mergeJobGroup)
self.mergeJobC.create(mergeJobGroup)
self.mainCleanupSub = Subscription(fileset=unmergedOutputFileset,
workflow=mainCleanupWorkflow,
type="Cleanup")
self.mainCleanupSub.create()
self.mainCleanupSub.acquireFiles(unmergedFileA)
self.mainCleanupSub.completeFiles(unmergedFileB)
cleanupJobGroup = JobGroup(subscription=self.mainCleanupSub)
cleanupJobGroup.create()
self.cleanupJobA = Job(name="CleanupJobA")
self.cleanupJobA["state"] = "new"
self.cleanupJobA["location"] = "site1"
self.cleanupJobB = Job(name="CleanupJobB")
self.cleanupJobB["state"] = "executing"
self.cleanupJobB["location"] = "site1"
self.cleanupJobC = Job(name="CleanupJobC")
self.cleanupJobC["state"] = "complete"
self.cleanupJobC["location"] = "site1"
self.cleanupJobA.create(cleanupJobGroup)
self.cleanupJobB.create(cleanupJobGroup)
self.cleanupJobC.create(cleanupJobGroup)
jobList = [
self.procJobA, self.procJobB, self.procJobC, self.mergeJobA,
self.mergeJobB, self.mergeJobC, self.cleanupJobA, self.cleanupJobB,
self.cleanupJobC
]
changeStateAction.execute(jobList)
if baAPI:
for job in jobList:
job['plugin'] = 'TestPlugin'
job['userdn'] = userDN
job['usergroup'] = 'DEFAULT'
job['userrole'] = 'DEFAULT'
job['custom']['location'] = 'site1'
baAPI.createNewJobs(wmbsJobs=jobList)
# We'll create an unrelated workflow to verify that it isn't affected
# by the killing code.
bogusFileset = Fileset("dontkillme")
bogusFileset.create()
bogusFileA = File("bogus/lfnA", locations="goodse.cern.ch")
bogusFileA.create()
bogusFileset.addFile(bogusFileA)
bogusFileset.commit()
bogusWorkflow = Workflow(spec="spec2",
owner="Steve",
name="Bogus",
task="Proc")
bogusWorkflow.create()
self.bogusSub = Subscription(fileset=bogusFileset,
workflow=bogusWorkflow,
type="Processing")
self.bogusSub.create()
self.bogusSub.acquireFiles(bogusFileA)
return
def verifyFileKillStatus(self):
"""
_verifyFileKillStatus_
Verify that all files were killed correctly. The status of files in
Cleanup and LogCollect subscriptions isn't modified. Status of
already completed and failed files is not modified. Also verify that
the bogus subscription is untouched.
"""
failedFiles = self.mainProcSub.filesOfStatus("Failed")
acquiredFiles = self.mainProcSub.filesOfStatus("Acquired")
completedFiles = self.mainProcSub.filesOfStatus("Completed")
availableFiles = self.mainProcSub.filesOfStatus("Available")
bogusAcquiredFiles = self.bogusSub.filesOfStatus("Acquired")
self.assertEqual(len(availableFiles), 0, \
"Error: There should be no available files.")
self.assertEqual(len(acquiredFiles), 0, \
"Error: There should be no acquired files.")
self.assertEqual(len(bogusAcquiredFiles), 1, \
"Error: There should be one acquired file.")
self.assertEqual(len(completedFiles), 3, \
"Error: There should be only one completed file.")
goldenLFNs = ["lfnA", "lfnB", "lfnC"]
for completedFile in completedFiles:
self.assertTrue(completedFile["lfn"] in goldenLFNs, \
"Error: Extra completed file.")
goldenLFNs.remove(completedFile["lfn"])
self.assertEqual(len(failedFiles), 0, \
"Error: There should be no failed files.")
self.assertEqual(len(goldenLFNs), 0, \
"Error: Missing LFN")
failedFiles = self.mainMergeSub.filesOfStatus("Failed")
acquiredFiles = self.mainMergeSub.filesOfStatus("Acquired")
completedFiles = self.mainMergeSub.filesOfStatus("Completed")
availableFiles = self.mainMergeSub.filesOfStatus("Available")
self.assertEqual(len(acquiredFiles), 0, \
"Error: Merge subscription should have 0 acq files.")
self.assertEqual(len(availableFiles), 0, \
"Error: Merge subscription should have 0 avail files.")
self.assertEqual(len(failedFiles), 1, \
"Error: Merge subscription should have 1 failed files.")
self.assertEqual(
list(failedFiles)[0]["lfn"], "ulfnB", "Error: Wrong failed file.")
self.assertEqual(len(completedFiles), 2, \
"Error: Merge subscription should have 2 compl files.")
goldenLFNs = ["ulfnA", "ulfnC"]
for completedFile in completedFiles:
self.assertTrue(completedFile["lfn"] in goldenLFNs, \
"Error: Extra complete file.")
goldenLFNs.remove(completedFile["lfn"])
self.assertEqual(len(goldenLFNs), 0, \
"Error: Missing LFN")
failedFiles = self.mainCleanupSub.filesOfStatus("Failed")
acquiredFiles = self.mainCleanupSub.filesOfStatus("Acquired")
completedFiles = self.mainCleanupSub.filesOfStatus("Completed")
availableFiles = self.mainCleanupSub.filesOfStatus("Available")
self.assertEqual(len(failedFiles), 0, \
"Error: Cleanup subscription should have 0 fai files.")
self.assertEqual(len(acquiredFiles), 1, \
"Error: There should be only one acquired file.")
self.assertEqual(list(acquiredFiles)[0]["lfn"], "ulfnA", \
"Error: Wrong acquired LFN.")
self.assertEqual(len(completedFiles), 1, \
"Error: There should be only one completed file.")
self.assertEqual(list(completedFiles)[0]["lfn"], "ulfnB", \
"Error: Wrong completed LFN.")
self.assertEqual(len(availableFiles), 1, \
"Error: There should be only one available file.")
self.assertEqual(list(availableFiles)[0]["lfn"], "ulfnC", \
"Error: Wrong completed LFN.")
return
def verifyJobKillStatus(self):
"""
_verifyJobKillStatus_
Verify that jobs are killed correctly. Jobs belonging to Cleanup and
LogCollect subscriptions are not killed. The status of jobs that have
already finished running is not changed.
"""
self.procJobA.load()
self.procJobB.load()
self.procJobC.load()
self.assertEqual(self.procJobA["state"], "killed", \
"Error: Proc job A should be killed.")
self.assertEqual(self.procJobB["state"], "killed", \
"Error: Proc job B should be killed.")
self.assertEqual(self.procJobC["state"], "complete", \
"Error: Proc job C should be complete.")
self.mergeJobA.load()
self.mergeJobB.load()
self.mergeJobC.load()
self.assertEqual(self.mergeJobA["state"], "exhausted", \
"Error: Merge job A should be exhausted.")
self.assertEqual(self.mergeJobB["state"], "cleanout", \
"Error: Merge job B should be cleanout.")
self.assertEqual(self.mergeJobC["state"], "killed", \
"Error: Merge job C should be killed.")
self.cleanupJobA.load()
self.cleanupJobB.load()
self.cleanupJobC.load()
self.assertEqual(self.cleanupJobA["state"], "new", \
"Error: Cleanup job A should be new.")
self.assertEqual(self.cleanupJobB["state"], "executing", \
"Error: Cleanup job B should be executing.")
self.assertEqual(self.cleanupJobC["state"], "complete", \
"Error: Cleanup job C should be complete.")
return
def createTestWMSpec(self):
"""
_createTestWMSpec_
Create a WMSpec that has a processing, merge, cleanup and skims tasks that
can be used by the subscription creation test.
"""
testWorkload = WMWorkloadHelper(WMWorkload("TestWorkload"))
testWorkload.setDashboardActivity("TestReReco")
testWorkload.setSpecUrl("/path/to/workload")
testWorkload.setOwnerDetails("sfoulkes", "DMWM", {'dn': 'MyDN'})
procTask = testWorkload.newTask("ProcessingTask")
procTask.setTaskType("Processing")
procTask.setSplittingAlgorithm("FileBased", files_per_job=1)
procTaskCMSSW = procTask.makeStep("cmsRun1")
procTaskCMSSW.setStepType("CMSSW")
procTaskCMSSWHelper = procTaskCMSSW.getTypeHelper()
procTask.setTaskType("Processing")
procTask.setSiteWhitelist(["site1"])
procTask.setSiteBlacklist(["site2"])
procTask.applyTemplates()
procTaskCMSSWHelper.addOutputModule("OutputA",
primaryDataset="bogusPrimary",
processedDataset="bogusProcessed",
dataTier="DataTierA",
lfnBase="bogusUnmerged",
mergedLFNBase="bogusMerged",
filterName=None)
mergeTask = procTask.addTask("MergeTask")
mergeTask.setInputReference(procTaskCMSSW, outputModule="OutputA")
mergeTask.setTaskType("Merge")
mergeTask.setSplittingAlgorithm("WMBSMergeBySize",
min_merge_size=1,
max_merge_size=2,
max_merge_events=3)
mergeTaskCMSSW = mergeTask.makeStep("cmsRun1")
mergeTaskCMSSW.setStepType("CMSSW")
mergeTaskCMSSWHelper = mergeTaskCMSSW.getTypeHelper()
mergeTask.setTaskType("Merge")
mergeTask.applyTemplates()
mergeTaskCMSSWHelper.addOutputModule("Merged",
primaryDataset="bogusPrimary",
processedDataset="bogusProcessed",
dataTier="DataTierA",
lfnBase="bogusUnmerged",
mergedLFNBase="bogusMerged",
filterName=None)
cleanupTask = procTask.addTask("CleanupTask")
cleanupTask.setInputReference(procTaskCMSSW, outputModule="OutputA")
cleanupTask.setTaskType("Merge")
cleanupTask.setSplittingAlgorithm("SiblingProcessingBased",
files_per_job=50)
cleanupTaskCMSSW = cleanupTask.makeStep("cmsRun1")
cleanupTaskCMSSW.setStepType("CMSSW")
cleanupTaskCMSSWHelper = cleanupTaskCMSSW.getTypeHelper()
cleanupTask.setTaskType("Cleanup")
cleanupTask.applyTemplates()
skimTask = mergeTask.addTask("SkimTask")
skimTask.setTaskType("Skim")
skimTask.setInputReference(mergeTaskCMSSW, outputModule="Merged")
skimTask.setSplittingAlgorithm("FileBased",
files_per_job=1,
include_parents=True)
skimTaskCMSSW = skimTask.makeStep("cmsRun1")
skimTaskCMSSW.setStepType("CMSSW")
skimTaskCMSSWHelper = skimTaskCMSSW.getTypeHelper()
skimTask.setTaskType("Skim")
skimTask.applyTemplates()
skimTaskCMSSWHelper.addOutputModule("SkimOutputA",
primaryDataset="bogusPrimary",
processedDataset="bogusProcessed",
dataTier="DataTierA",
lfnBase="bogusUnmerged",
mergedLFNBase="bogusMerged",
filterName=None)
skimTaskCMSSWHelper.addOutputModule("SkimOutputB",
primaryDataset="bogusPrimary",
processedDataset="bogusProcessed",
dataTier="DataTierA",
lfnBase="bogusUnmerged",
mergedLFNBase="bogusMerged",
filterName=None)
return testWorkload
def setupMCWMSpec(self):
"""Setup MC workflow"""
self.wmspec = self.createMCWMSpec()
self.topLevelTask = getFirstTask(self.wmspec)
self.inputDataset = self.topLevelTask.inputDataset()
self.dataset = self.topLevelTask.getInputDatasetPath()
self.dbs = None
self.siteDB = fakeSiteDB()
# add sites that would normally be added by operator via resource_control
locationDAO = self.daoFactory(classname="Locations.New")
self.ses = []
for site in ['T2_XX_SiteA', 'T2_XX_SiteB']:
locationDAO.execute(siteName=site,
seName=self.siteDB.cmsNametoSE(site)[0])
self.ses.append(self.siteDB.cmsNametoSE(site)[0])
def createWMSpec(self, name='ReRecoWorkload'):
factory = ReRecoWorkloadFactory()
rerecoArgs["ConfigCacheID"] = createConfig(rerecoArgs["CouchDBName"])
wmspec = factory.factoryWorkloadConstruction(name, rerecoArgs)
wmspec.setSpecUrl("/path/to/workload")
wmspec.setSubscriptionInformation(custodialSites=[],
nonCustodialSites=[],
autoApproveSites=[],
priority="Low",
custodialSubType="Move")
return wmspec
def createMCWMSpec(self, name='MonteCarloWorkload'):
wmspec = monteCarloWorkload(name, mcArgs)
wmspec.setSpecUrl("/path/to/workload")
getFirstTask(wmspec).addProduction(totalevents=10000)
return wmspec
def getDBS(self, wmspec):
topLevelTask = getFirstTask(wmspec)
inputDataset = topLevelTask.inputDataset()
dbs = MockDBSReader(inputDataset.dbsurl)
#dbsDict = {self.inputDataset.dbsurl : self.dbs}
return dbs
def createWMBSHelperWithTopTask(self,
wmspec,
block,
mask=None,
parentFlag=False,
detail=False):
topLevelTask = getFirstTask(wmspec)
wmbs = WMBSHelper(wmspec,
topLevelTask.name(),
block,
mask,
cachepath=self.workDir)
if block:
if parentFlag:
block = self.dbs.getFileBlockWithParents(block)[block]
else:
block = self.dbs.getFileBlock(block)[block]
sub, files = wmbs.createSubscriptionAndAddFiles(block=block)
if detail:
return wmbs, sub, files
else:
return wmbs
def testKillWorkflow(self):
"""
_testKillWorkflow_
Verify that workflow killing works correctly.
"""
configFile = EmulatorSetup.setupWMAgentConfig()
config = loadConfigurationFile(configFile)
baAPI = BossAirAPI(config=config)
# Create nine jobs
self.setupForKillTest(baAPI=baAPI)
self.assertEqual(len(baAPI._listRunJobs()), 9)
killWorkflow("Main", config, config)
self.verifyFileKillStatus()
self.verifyJobKillStatus()
self.assertEqual(len(baAPI._listRunJobs()), 8)
EmulatorSetup.deleteConfig(configFile)
return
def testCreateSubscription(self):
"""
_testCreateSubscription_
Verify that the subscription creation code works correctly.
"""
resourceControl = ResourceControl()
resourceControl.insertSite(siteName='site1',
seName='goodse.cern.ch',
ceName='site1',
plugin="TestPlugin")
resourceControl.insertSite(siteName='site2',
seName='goodse2.cern.ch',
ceName='site2',
plugin="TestPlugin")
testWorkload = self.createTestWMSpec()
testTopLevelTask = getFirstTask(testWorkload)
testWMBSHelper = WMBSHelper(testWorkload,
testTopLevelTask.name(),
"SomeBlock",
cachepath=self.workDir)
testWMBSHelper.createTopLevelFileset()
testWMBSHelper._createSubscriptionsInWMBS(
testTopLevelTask, testWMBSHelper.topLevelFileset)
procWorkflow = Workflow(name="TestWorkload",
task="/TestWorkload/ProcessingTask")
procWorkflow.load()
self.assertEqual(procWorkflow.owner, "sfoulkes",
"Error: Wrong owner: %s" % procWorkflow.owner)
self.assertEqual(procWorkflow.group, "DMWM",
"Error: Wrong group: %s" % procWorkflow.group)
self.assertEqual(procWorkflow.wfType, "TestReReco",
"Error: Wrong type.")
self.assertEqual(
procWorkflow.spec,
os.path.join(self.workDir, procWorkflow.name, "WMSandbox",
"WMWorkload.pkl"), "Error: Wrong spec URL")
self.assertEqual(len(procWorkflow.outputMap.keys()), 1,
"Error: Wrong number of WF outputs.")
mergedProcOutput = procWorkflow.outputMap["OutputA"][0][
"merged_output_fileset"]
unmergedProcOutput = procWorkflow.outputMap["OutputA"][0][
"output_fileset"]
mergedProcOutput.loadData()
unmergedProcOutput.loadData()
self.assertEqual(
mergedProcOutput.name,
"/TestWorkload/ProcessingTask/MergeTask/merged-Merged",
"Error: Merged output fileset is wrong.")
self.assertEqual(unmergedProcOutput.name,
"/TestWorkload/ProcessingTask/unmerged-OutputA",
"Error: Unmerged output fileset is wrong.")
mergeWorkflow = Workflow(name="TestWorkload",
task="/TestWorkload/ProcessingTask/MergeTask")
mergeWorkflow.load()
self.assertEqual(mergeWorkflow.owner, "sfoulkes",
"Error: Wrong owner.")
self.assertEqual(
mergeWorkflow.spec,
os.path.join(self.workDir, mergeWorkflow.name, "WMSandbox",
"WMWorkload.pkl"), "Error: Wrong spec URL")
self.assertEqual(len(mergeWorkflow.outputMap.keys()), 1,
"Error: Wrong number of WF outputs.")
cleanupWorkflow = Workflow(
name="TestWorkload",
task="/TestWorkload/ProcessingTask/CleanupTask")
cleanupWorkflow.load()
self.assertEqual(cleanupWorkflow.owner, "sfoulkes",
"Error: Wrong owner.")
self.assertEqual(
cleanupWorkflow.spec,
os.path.join(self.workDir, cleanupWorkflow.name, "WMSandbox",
"WMWorkload.pkl"), "Error: Wrong spec URL")
self.assertEqual(len(cleanupWorkflow.outputMap.keys()), 0,
"Error: Wrong number of WF outputs.")
unmergedMergeOutput = mergeWorkflow.outputMap["Merged"][0][
"output_fileset"]
unmergedMergeOutput.loadData()
self.assertEqual(
unmergedMergeOutput.name,
"/TestWorkload/ProcessingTask/MergeTask/merged-Merged",
"Error: Unmerged output fileset is wrong.")
skimWorkflow = Workflow(
name="TestWorkload",
task="/TestWorkload/ProcessingTask/MergeTask/SkimTask")
skimWorkflow.load()
self.assertEqual(skimWorkflow.owner, "sfoulkes", "Error: Wrong owner.")
self.assertEqual(
skimWorkflow.spec,
os.path.join(self.workDir, skimWorkflow.name, "WMSandbox",
"WMWorkload.pkl"), "Error: Wrong spec URL")
self.assertEqual(len(skimWorkflow.outputMap.keys()), 2,
"Error: Wrong number of WF outputs.")
mergedSkimOutputA = skimWorkflow.outputMap["SkimOutputA"][0][
"merged_output_fileset"]
unmergedSkimOutputA = skimWorkflow.outputMap["SkimOutputA"][0][
"output_fileset"]
mergedSkimOutputB = skimWorkflow.outputMap["SkimOutputB"][0][
"merged_output_fileset"]
unmergedSkimOutputB = skimWorkflow.outputMap["SkimOutputB"][0][
"output_fileset"]
mergedSkimOutputA.loadData()
mergedSkimOutputB.loadData()
unmergedSkimOutputA.loadData()
unmergedSkimOutputB.loadData()
self.assertEqual(
mergedSkimOutputA.name,
"/TestWorkload/ProcessingTask/MergeTask/SkimTask/unmerged-SkimOutputA",
"Error: Merged output fileset is wrong: %s" %
mergedSkimOutputA.name)
self.assertEqual(
unmergedSkimOutputA.name,
"/TestWorkload/ProcessingTask/MergeTask/SkimTask/unmerged-SkimOutputA",
"Error: Unmerged output fileset is wrong.")
self.assertEqual(
mergedSkimOutputB.name,
"/TestWorkload/ProcessingTask/MergeTask/SkimTask/unmerged-SkimOutputB",
"Error: Merged output fileset is wrong.")
self.assertEqual(
unmergedSkimOutputB.name,
"/TestWorkload/ProcessingTask/MergeTask/SkimTask/unmerged-SkimOutputB",
"Error: Unmerged output fileset is wrong.")
topLevelFileset = Fileset(name="TestWorkload-ProcessingTask-SomeBlock")
topLevelFileset.loadData()
procSubscription = Subscription(fileset=topLevelFileset,
workflow=procWorkflow)
procSubscription.loadData()
self.assertEqual(len(procSubscription.getWhiteBlackList()), 2,
"Error: Wrong site white/black list for proc sub.")
for site in procSubscription.getWhiteBlackList():
if site["site_name"] == "site1":
self.assertEqual(site["valid"], 1,
"Error: Site should be white listed.")
else:
self.assertEqual(site["valid"], 0,
"Error: Site should be black listed.")
self.assertEqual(procSubscription["type"], "Processing",
"Error: Wrong subscription type.")
self.assertEqual(procSubscription["split_algo"], "FileBased",
"Error: Wrong split algo.")
mergeSubscription = Subscription(fileset=unmergedProcOutput,
workflow=mergeWorkflow)
mergeSubscription.loadData()
self.assertEqual(len(mergeSubscription.getWhiteBlackList()), 0,
"Error: Wrong white/black list for merge sub.")
self.assertEqual(mergeSubscription["type"], "Merge",
"Error: Wrong subscription type.")
self.assertEqual(mergeSubscription["split_algo"], "WMBSMergeBySize",
"Error: Wrong split algo.")
skimSubscription = Subscription(fileset=unmergedMergeOutput,
workflow=skimWorkflow)
skimSubscription.loadData()
self.assertEqual(skimSubscription["type"], "Skim",
"Error: Wrong subscription type.")
self.assertEqual(skimSubscription["split_algo"], "FileBased",
"Error: Wrong split algo.")
return
def testTruncatedWFInsertion(self):
"""
_testTruncatedWFInsertion_
"""
resourceControl = ResourceControl()
resourceControl.insertSite(siteName='site1',
seName='goodse.cern.ch',
ceName='site1',
plugin="TestPlugin")
resourceControl.insertSite(siteName='site2',
seName='goodse2.cern.ch',
ceName='site2',
plugin="TestPlugin")
testWorkload = self.createTestWMSpec()
testTopLevelTask = getFirstTask(testWorkload)
testWMBSHelper = WMBSHelper(testWorkload,
testTopLevelTask.name(),
"SomeBlock",
cachepath=self.workDir)
testWMBSHelper.createTopLevelFileset()
testWMBSHelper._createSubscriptionsInWMBS(
testTopLevelTask, testWMBSHelper.topLevelFileset)
testWorkload.truncate("ResubmitTestWorkload",
"/TestWorkload/ProcessingTask/MergeTask",
"someserver", "somedatabase")
# create the subscription for multiple top task (MergeTask and CleanupTask for the same block)
for task in testWorkload.getTopLevelTask():
testResubmitWMBSHelper = WMBSHelper(testWorkload,
task.name(),
"SomeBlock2",
cachepath=self.workDir)
testResubmitWMBSHelper.createTopLevelFileset()
testResubmitWMBSHelper._createSubscriptionsInWMBS(
task, testResubmitWMBSHelper.topLevelFileset)
mergeWorkflow = Workflow(name="ResubmitTestWorkload",
task="/ResubmitTestWorkload/MergeTask")
mergeWorkflow.load()
self.assertEqual(mergeWorkflow.owner, "sfoulkes",
"Error: Wrong owner.")
self.assertEqual(
mergeWorkflow.spec,
os.path.join(self.workDir, mergeWorkflow.name, "WMSandbox",
"WMWorkload.pkl"), "Error: Wrong spec URL")
self.assertEqual(len(mergeWorkflow.outputMap.keys()), 1,
"Error: Wrong number of WF outputs.")
unmergedMergeOutput = mergeWorkflow.outputMap["Merged"][0][
"output_fileset"]
unmergedMergeOutput.loadData()
self.assertEqual(unmergedMergeOutput.name,
"/ResubmitTestWorkload/MergeTask/merged-Merged",
"Error: Unmerged output fileset is wrong.")
skimWorkflow = Workflow(
name="ResubmitTestWorkload",
task="/ResubmitTestWorkload/MergeTask/SkimTask")
skimWorkflow.load()
self.assertEqual(skimWorkflow.owner, "sfoulkes", "Error: Wrong owner.")
self.assertEqual(
skimWorkflow.spec,
os.path.join(self.workDir, skimWorkflow.name, "WMSandbox",
"WMWorkload.pkl"), "Error: Wrong spec URL")
self.assertEqual(len(skimWorkflow.outputMap.keys()), 2,
"Error: Wrong number of WF outputs.")
mergedSkimOutputA = skimWorkflow.outputMap["SkimOutputA"][0][
"merged_output_fileset"]
unmergedSkimOutputA = skimWorkflow.outputMap["SkimOutputA"][0][
"output_fileset"]
mergedSkimOutputB = skimWorkflow.outputMap["SkimOutputB"][0][
"merged_output_fileset"]
unmergedSkimOutputB = skimWorkflow.outputMap["SkimOutputB"][0][
"output_fileset"]
mergedSkimOutputA.loadData()
mergedSkimOutputB.loadData()
unmergedSkimOutputA.loadData()
unmergedSkimOutputB.loadData()
self.assertEqual(
mergedSkimOutputA.name,
"/ResubmitTestWorkload/MergeTask/SkimTask/unmerged-SkimOutputA",
"Error: Merged output fileset is wrong: %s" %
mergedSkimOutputA.name)
self.assertEqual(
unmergedSkimOutputA.name,
"/ResubmitTestWorkload/MergeTask/SkimTask/unmerged-SkimOutputA",
"Error: Unmerged output fileset is wrong.")
self.assertEqual(
mergedSkimOutputB.name,
"/ResubmitTestWorkload/MergeTask/SkimTask/unmerged-SkimOutputB",
"Error: Merged output fileset is wrong.")
self.assertEqual(
unmergedSkimOutputB.name,
"/ResubmitTestWorkload/MergeTask/SkimTask/unmerged-SkimOutputB",
"Error: Unmerged output fileset is wrong.")
topLevelFileset = Fileset(
name="ResubmitTestWorkload-MergeTask-SomeBlock2")
topLevelFileset.loadData()
mergeSubscription = Subscription(fileset=topLevelFileset,
workflow=mergeWorkflow)
mergeSubscription.loadData()
self.assertEqual(len(mergeSubscription.getWhiteBlackList()), 0,
"Error: Wrong white/black list for merge sub.")
self.assertEqual(mergeSubscription["type"], "Merge",
"Error: Wrong subscription type.")
self.assertEqual(mergeSubscription["split_algo"], "WMBSMergeBySize",
"Error: Wrong split algo.")
skimSubscription = Subscription(fileset=unmergedMergeOutput,
workflow=skimWorkflow)
skimSubscription.loadData()
self.assertEqual(skimSubscription["type"], "Skim",
"Error: Wrong subscription type.")
self.assertEqual(skimSubscription["split_algo"], "FileBased",
"Error: Wrong split algo.")
return
def testReReco(self):
"""ReReco workflow"""
# create workflow
block = self.dataset + "#1"
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
files = wmbs.validFiles(self.dbs.getFileBlock(block))
self.assertEqual(len(files), 1)
def testReRecoBlackRunRestriction(self):
"""ReReco workflow with Run restrictions"""
block = self.dataset + "#2"
#add run blacklist
self.topLevelTask.setInputRunBlacklist([1, 2, 3, 4])
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
files = wmbs.validFiles(self.dbs.getFileBlock(block)[block]['Files'])
self.assertEqual(len(files), 0)
def testReRecoWhiteRunRestriction(self):
block = self.dataset + "#2"
# Run Whitelist
self.topLevelTask.setInputRunWhitelist([2])
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
files = wmbs.validFiles(self.dbs.getFileBlock(block)[block]['Files'])
self.assertEqual(len(files), GlobalParams.numOfFilesPerBlock())
def testLumiMaskRestrictionsOK(self):
block = self.dataset + "#1"
self.wmspec.getTopLevelTask()[0].data.input.splitting.runs = ['1']
self.wmspec.getTopLevelTask()[0].data.input.splitting.lumis = ['1,1']
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
files = wmbs.validFiles(self.dbs.getFileBlock(block)[block]['Files'])
self.assertEqual(len(files), GlobalParams.numOfFilesPerBlock())
def testLumiMaskRestrictionsKO(self):
block = self.dataset + "#1"
self.wmspec.getTopLevelTask()[0].data.input.splitting.runs = [
'123454321'
]
self.wmspec.getTopLevelTask()[0].data.input.splitting.lumis = [
'123,123'
]
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
files = wmbs.validFiles(self.dbs.getFileBlock(block)[block]['Files'])
self.assertEqual(len(files), 0)
def testDuplicateFileInsert(self):
# using default wmspec
block = self.dataset + "#1"
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
wmbs.topLevelFileset.loadData()
numOfFiles = len(wmbs.topLevelFileset.files)
# check initially inserted files.
dbsFiles = self.dbs.getFileBlock(block)[block]['Files']
self.assertEqual(numOfFiles, len(dbsFiles))
firstFileset = wmbs.topLevelFileset
wmbsDao = wmbs.daofactory(classname="Files.InFileset")
numOfFiles = len(wmbsDao.execute(firstFileset.id))
self.assertEqual(numOfFiles, len(dbsFiles))
# use the new spec with same inputdataset
block = self.dataset + "#1"
wmspec = self.createWMSpec("TestSpec1")
dbs = self.getDBS(wmspec)
wmbs = self.createWMBSHelperWithTopTask(wmspec, block)
# check duplicate insert
dbsFiles = dbs.getFileBlock(block)[block]['Files']
numOfFiles = wmbs.addFiles(dbs.getFileBlock(block)[block])
self.assertEqual(numOfFiles, 0)
secondFileset = wmbs.topLevelFileset
wmbsDao = wmbs.daofactory(classname="Files.InFileset")
numOfFiles = len(wmbsDao.execute(secondFileset.id))
self.assertEqual(numOfFiles, len(dbsFiles))
self.assertNotEqual(firstFileset.id, secondFileset.id)
def testDuplicateSubscription(self):
"""Can't duplicate subscriptions"""
# using default wmspec
block = self.dataset + "#1"
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
wmbs.topLevelFileset.loadData()
numOfFiles = len(wmbs.topLevelFileset.files)
filesetId = wmbs.topLevelFileset.id
subId = wmbs.topLevelSubscription['id']
# check initially inserted files.
dbsFiles = self.dbs.getFileBlock(block)[block]['Files']
self.assertEqual(numOfFiles, len(dbsFiles))
firstFileset = wmbs.topLevelFileset
self.assertEqual(numOfFiles, len(dbsFiles))
# reinsert subscription - shouldn't create anything new
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
wmbs.topLevelFileset.loadData()
self.assertEqual(numOfFiles, len(wmbs.topLevelFileset.files))
self.assertEqual(filesetId, wmbs.topLevelFileset.id)
self.assertEqual(subId, wmbs.topLevelSubscription['id'])
# now do a montecarlo workflow
self.setupMCWMSpec()
mask = Mask(FirstRun=12,
FirstLumi=1234,
FirstEvent=12345,
LastEvent=999995,
LastLumi=12345,
LastRun=12)
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, None, mask)
wmbs.topLevelFileset.loadData()
numOfFiles = len(wmbs.topLevelFileset.files)
filesetId = wmbs.topLevelFileset.id
subId = wmbs.topLevelSubscription['id']
# check initially inserted files.
numDbsFiles = 1
self.assertEqual(numOfFiles, numDbsFiles)
firstFileset = wmbs.topLevelFileset
self.assertEqual(numOfFiles, numDbsFiles)
# reinsert subscription - shouldn't create anything new
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, None, mask)
wmbs.topLevelFileset.loadData()
self.assertEqual(numOfFiles, len(wmbs.topLevelFileset.files))
self.assertEqual(filesetId, wmbs.topLevelFileset.id)
self.assertEqual(subId, wmbs.topLevelSubscription['id'])
def testParentage(self):
"""
1. check whether parent files are created in wmbs.
2. check parent files are associated to child.
3. When 2 specs with the same input data (one with parent processing, one without it)
is inserted, if one without parent processing inserted first then the other with
parent processing insert, it still needs to create parent files although child files
are duplicate
"""
block = self.dataset + "#1"
wmbs, sub, numFiles = self.createWMBSHelperWithTopTask(
self.wmspec, block, parentFlag=False, detail=True)
# file creation without parents
self.assertEqual(GlobalParams.numOfFilesPerBlock(), numFiles)
wmbs.topLevelFileset.loadData()
for child in wmbs.topLevelFileset.files:
# no parent per child
self.assertEqual(len(child["parents"]), 0)
wmbs, sub, numFiles = self.createWMBSHelperWithTopTask(self.wmspec,
block,
parentFlag=True,
detail=True)
self.assertEqual(GlobalParams.numOfFilesPerBlock(), numFiles)
wmbs.topLevelFileset.loadData()
for child in wmbs.topLevelFileset.files:
# one parent per child
self.assertEqual(len(child["parents"]), 1)
def testMCFakeFileInjection(self):
"""Inject fake Monte Carlo files into WMBS"""
self.setupMCWMSpec()
mask = Mask(FirstRun=12,
FirstLumi=1234,
FirstEvent=12345,
LastEvent=999995,
LastLumi=12345,
LastRun=12)
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, None, mask)
subscription = wmbs.topLevelSubscription
self.assertEqual(1, subscription.exists())
fileset = subscription['fileset']
self.assertEqual(1, fileset.exists())
fileset.loadData() # need to refresh from database
self.assertEqual(len(fileset.files), 1)
self.assertEqual(len(fileset.parents), 0)
self.assertFalse(fileset.open)
file = list(fileset.files)[0]
self.assertEqual(file['events'], mask['LastEvent'] -
mask['FirstEvent'] + 1) # inclusive range
self.assertEqual(file['merged'],
False) # merged files get added to dbs
self.assertEqual(len(file['parents']), 0)
#file.loadData()
self.assertEqual(sorted(file['locations']), sorted(self.ses))
self.assertEqual(len(file.getParentLFNs()), 0)
self.assertEqual(len(file.getRuns()), 1)
run = file.getRuns()[0]
self.assertEqual(run.run, mask['FirstRun'])
self.assertEqual(run.lumis[0], mask['FirstLumi'])
self.assertEqual(run.lumis[-1], mask['LastLumi'])
self.assertEqual(len(run.lumis),
mask['LastLumi'] - mask['FirstLumi'] + 1)
def testGetOutputMapDAO(self):
"""
_testGetOutputMapDAO_
Verify the proper behavior of the GetOutputMapDAO for a variety of
different processing chains.
"""
recoOutputFileset = Fileset(name="RECO")
recoOutputFileset.create()
mergedRecoOutputFileset = Fileset(name="MergedRECO")
mergedRecoOutputFileset.create()
alcaOutputFileset = Fileset(name="ALCA")
alcaOutputFileset.create()
mergedAlcaOutputFileset = Fileset(name="MergedALCA")
mergedAlcaOutputFileset.create()
dqmOutputFileset = Fileset(name="DQM")
dqmOutputFileset.create()
mergedDqmOutputFileset = Fileset(name="MergedDQM")
mergedDqmOutputFileset.create()
cleanupFileset = Fileset(name="Cleanup")
cleanupFileset.create()
testWorkflow = Workflow(spec="wf001.xml",
owner="Steve",
name="TestWF",
task="None")
testWorkflow.create()
testWorkflow.addOutput("output", recoOutputFileset,
mergedRecoOutputFileset)
testWorkflow.addOutput("ALCARECOStreamCombined", alcaOutputFileset,
mergedAlcaOutputFileset)
testWorkflow.addOutput("DQM", dqmOutputFileset, mergedDqmOutputFileset)
testWorkflow.addOutput("output", cleanupFileset)
testWorkflow.addOutput("ALCARECOStreamCombined", cleanupFileset)
testWorkflow.addOutput("DQM", cleanupFileset)
testRecoMergeWorkflow = Workflow(spec="wf002.xml",
owner="Steve",
name="TestRecoMergeWF",
task="None")
testRecoMergeWorkflow.create()
testRecoMergeWorkflow.addOutput("anything", mergedRecoOutputFileset,
mergedRecoOutputFileset)
testRecoProcWorkflow = Workflow(spec="wf004.xml",
owner="Steve",
name="TestRecoProcWF",
task="None")
testRecoProcWorkflow.create()
testAlcaChildWorkflow = Workflow(spec="wf003.xml",
owner="Steve",
name="TestAlcaChildWF",
task="None")
testAlcaChildWorkflow.create()
inputFile = File(lfn="/path/to/some/lfn",
size=600000,
events=60000,
locations="cmssrm.fnal.gov")
inputFile.create()
testFileset = Fileset(name="TestFileset")
testFileset.create()
testFileset.addFile(inputFile)
testFileset.commit()
testSubscription = Subscription(fileset=testFileset,
workflow=testWorkflow,
split_algo="EventBased",
type="Processing")
testMergeRecoSubscription = Subscription(
fileset=recoOutputFileset,
workflow=testRecoMergeWorkflow,
split_algo="WMBSMergeBySize",
type="Merge")
testProcRecoSubscription = Subscription(fileset=recoOutputFileset,
workflow=testRecoProcWorkflow,
split_algo="FileBased",
type="Processing")
testChildAlcaSubscription = Subscription(
fileset=alcaOutputFileset,
workflow=testAlcaChildWorkflow,
split_algo="FileBased",
type="Processing")
testSubscription.create()
testMergeRecoSubscription.create()
testProcRecoSubscription.create()
testChildAlcaSubscription.create()
testSubscription.acquireFiles()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testJob = Job(name="SplitJobA", files=[inputFile])
testJob.create(group=testJobGroup)
testJob["state"] = "complete"
testJob.save()
outputMapAction = self.daoFactory(classname="Jobs.GetOutputMap")
outputMap = outputMapAction.execute(jobID=testJob["id"])
assert len(outputMap.keys()) == 3, \
"Error: Wrong number of outputs for primary workflow."
goldenMap = {
"output": (recoOutputFileset.id, mergedRecoOutputFileset.id),
"ALCARECOStreamCombined":
(alcaOutputFileset.id, mergedAlcaOutputFileset.id),
"DQM": (dqmOutputFileset.id, mergedDqmOutputFileset.id)
}
for outputID in outputMap.keys():
for outputFilesets in outputMap[outputID]:
if outputFilesets["merged_output_fileset"] == None:
self.assertEqual(outputFilesets["output_fileset"],
cleanupFileset.id,
"Error: Cleanup fileset is wrong.")
continue
self.assertTrue(outputID in goldenMap.keys(),
"Error: Output identifier is missing.")
self.assertEqual(outputFilesets["output_fileset"],
goldenMap[outputID][0],
"Error: Output fileset is wrong.")
self.assertEqual(outputFilesets["merged_output_fileset"],
goldenMap[outputID][1],
"Error: Merged output fileset is wrong.")
del goldenMap[outputID]
self.assertEqual(len(goldenMap.keys()), 0,
"Error: Missing output maps.")
return
def createJobs(self):
"""
_createJobs_
Create test jobs in WMBS and BossAir
"""
testWorkflow = Workflow(spec=makeUUID(), owner="tapas",
name=makeUUID(), task="Test")
testWorkflow.create()
testFilesetA = Fileset(name="TestFilesetA")
testFilesetA.create()
testFilesetB = Fileset(name="TestFilesetB")
testFilesetB.create()
testFilesetC = Fileset(name="TestFilesetC")
testFilesetC.create()
testFileA = File(lfn="testFileA", locations=set(["testSE1", "testSE2"]))
testFileA.create()
testFilesetA.addFile(testFileA)
testFilesetA.commit()
testFilesetB.addFile(testFileA)
testFilesetB.commit()
testFilesetC.addFile(testFileA)
testFilesetC.commit()
testSubscriptionA = Subscription(fileset=testFilesetA,
workflow=testWorkflow,
type="Processing")
testSubscriptionA.create()
testSubscriptionA.addWhiteBlackList([{"site_name": "testSite1", "valid": True}])
testSubscriptionB = Subscription(fileset=testFilesetB,
workflow=testWorkflow,
type="Processing")
testSubscriptionB.create()
testSubscriptionB.addWhiteBlackList([{"site_name": "testSite1", "valid": False}])
testSubscriptionC = Subscription(fileset=testFilesetC,
workflow=testWorkflow,
type="Merge")
testSubscriptionC.create()
testJobGroupA = JobGroup(subscription=testSubscriptionA)
testJobGroupA.create()
testJobGroupB = JobGroup(subscription=testSubscriptionB)
testJobGroupB.create()
testJobGroupC = JobGroup(subscription=testSubscriptionC)
testJobGroupC.create()
# Site1, Has been assigned a location and is complete.
testJobA = Job(name="testJobA", files=[testFileA])
testJobA["couch_record"] = makeUUID()
testJobA.create(group=testJobGroupA)
testJobA["state"] = "success"
# Site 1, Has been assigned a location and is incomplete.
testJobB = Job(name="testJobB", files=[testFileA])
testJobB["couch_record"] = makeUUID()
testJobB["cache_dir"] = self.tempDir
testJobB.create(group=testJobGroupA)
testJobB["state"] = "executing"
runJobB = RunJob()
runJobB.buildFromJob(testJobB)
runJobB["status"] = "PEND"
# Does not have a location, white listed to site 1
testJobC = Job(name="testJobC", files=[testFileA])
testJobC["couch_record"] = makeUUID()
testJobC.create(group=testJobGroupA)
testJobC["state"] = "new"
# Site 2, Has been assigned a location and is complete.
testJobD = Job(name="testJobD", files=[testFileA])
testJobD["couch_record"] = makeUUID()
testJobD.create(group=testJobGroupB)
testJobD["state"] = "success"
# Site 2, Has been assigned a location and is incomplete.
testJobE = Job(name="testJobE", files=[testFileA])
testJobE["couch_record"] = makeUUID()
testJobE.create(group=testJobGroupB)
testJobE["state"] = "executing"
runJobE = RunJob()
runJobE.buildFromJob(testJobE)
runJobE["status"] = "RUN"
# Does not have a location, site 1 is blacklisted.
testJobF = Job(name="testJobF", files=[testFileA])
testJobF["couch_record"] = makeUUID()
testJobF.create(group=testJobGroupB)
testJobF["state"] = "new"
# Site 3, Has been assigned a location and is complete.
testJobG = Job(name="testJobG", files=[testFileA])
testJobG["couch_record"] = makeUUID()
testJobG.create(group=testJobGroupC)
testJobG["state"] = "cleanout"
# Site 3, Has been assigned a location and is incomplete.
testJobH = Job(name="testJobH", files=[testFileA])
testJobH["couch_record"] = makeUUID()
testJobH.create(group=testJobGroupC)
testJobH["state"] = "new"
# Site 3, Does not have a location.
testJobI = Job(name="testJobI", files=[testFileA])
testJobI["couch_record"] = makeUUID()
testJobI.create(group=testJobGroupC)
testJobI["state"] = "new"
# Site 3, Does not have a location and is in cleanout.
testJobJ = Job(name="testJobJ", files=[testFileA])
testJobJ["couch_record"] = makeUUID()
testJobJ.create(group=testJobGroupC)
testJobJ["state"] = "cleanout"
changeStateAction = self.daoFactory(classname="Jobs.ChangeState")
changeStateAction.execute(jobs=[testJobA, testJobB, testJobC, testJobD,
testJobE, testJobF, testJobG, testJobH,
testJobI, testJobJ])
self.insertRunJob.execute([runJobB, runJobE])
setLocationAction = self.daoFactory(classname="Jobs.SetLocation")
setLocationAction.execute(testJobA["id"], "testSite1")
setLocationAction.execute(testJobB["id"], "testSite1")
setLocationAction.execute(testJobD["id"], "testSite1")
setLocationAction.execute(testJobE["id"], "testSite2")
setLocationAction.execute(testJobG["id"], "testSite1")
setLocationAction.execute(testJobH["id"], "testSite1")
return
def notestCreateDeleteExists(self):
"""
Create and then delete a job and workflow. Use the workunit class's exists() method to
determine if the workunit has been written to the database before the job is
created, after the job has been created, and after the workflow has been deleted.
"""
testWorkflow = Workflow(spec="spec.xml",
owner="Simon",
name="wf001",
task="Test")
testWorkflow.create()
testWMBSFileset = Fileset(name="TestFileset")
testWMBSFileset.create()
testSubscription = Subscription(fileset=testWMBSFileset,
workflow=testWorkflow)
testSubscription.create()
testJobGroup = JobGroup(subscription=testSubscription)
testJobGroup.create()
testFileA = File(lfn="/this/is/a/lfnA", size=1024, events=10)
testFileA.addRun(Run(1, *[45]))
testFileB = File(lfn="/this/is/a/lfnB", size=1024, events=10)
testFileB.addRun(Run(1, *[46]))
testFileA.create()
testFileB.create()
testJob = Job(name="TestJob", files=[testFileA, testFileB])
testWU1 = WorkUnit(taskID=testWorkflow.id,
fileid=testFileA['id'],
runLumi=Run(1, *[45]))
testWU2 = WorkUnit(taskID=testWorkflow.id,
fileid=testFileB['id'],
runLumi=Run(1, *[46]))
self.assertFalse(testWU1.exists(),
"WorkUnit exists before job was created")
self.assertFalse(testWU2.exists(),
"WorkUnit exists before job was created")
testJob.create(group=testJobGroup)
self.assertTrue(testWU1.exists(),
"WorkUnit does not exist after job was created")
self.assertTrue(testWU2.exists(),
"WorkUnit does not exist after job was created")
testJob.delete()
self.assertTrue(testWU1.exists(),
"WorkUnit does not exist after job is deleted")
self.assertTrue(testWU2.exists(),
"WorkUnit does not exist after job is deleted")
testWorkflow.delete()
self.assertFalse(testWU1.exists(),
"WorkUnit exists after workflow is deleted")
self.assertFalse(testWU2.exists(),
"WorkUnit exists after workflow is deleted")
return
class WMBSHelperTest(unittest.TestCase):
def setUp(self):
"""
_setUp_
"""
self.testInit = TestInitCouchApp(__file__)
self.testInit.setLogging()
self.testInit.setDatabaseConnection()
self.testInit.setupCouch("wmbshelper_t/jobs", "JobDump")
self.testInit.setupCouch("wmbshelper_t/fwjrs", "FWJRDump")
os.environ["COUCHDB"] = "wmbshelper_t"
self.testInit.setSchema(customModules = ["WMCore.WMBS",
"WMComponent.DBSBuffer.Database",
"WMCore.BossAir",
"WMCore.ResourceControl"],
useDefault = False)
self.workDir = self.testInit.generateWorkDir()
self.wmspec = self.createWMSpec()
self.topLevelTask = getFirstTask(self.wmspec)
self.inputDataset = self.topLevelTask.inputDataset()
self.dataset = self.topLevelTask.getInputDatasetPath()
self.dbs = MockDBSReader(self.inputDataset.dbsurl)
self.daoFactory = DAOFactory(package = "WMCore.WMBS",
logger = threading.currentThread().logger,
dbinterface = threading.currentThread().dbi)
return
def tearDown(self):
"""
_tearDown_
Clear out the database.
"""
self.testInit.clearDatabase()
self.testInit.tearDownCouch()
self.testInit.delWorkDir()
return
def setupForKillTest(self, baAPI = None):
"""
_setupForKillTest_
Inject a workflow into WMBS that has a processing task, a merge task and
a cleanup task. Inject files into the various tasks at various
processing states (acquired, complete, available...). Also create jobs
for each subscription in various states.
"""
myThread = threading.currentThread()
daoFactory = DAOFactory(package = "WMCore.WMBS",
logger = myThread.logger,
dbinterface = myThread.dbi)
locationAction = daoFactory(classname = "Locations.New")
changeStateAction = daoFactory(classname = "Jobs.ChangeState")
resourceControl = ResourceControl()
resourceControl.insertSite(siteName = 'site1', seName = 'goodse.cern.ch',
ceName = 'site1', plugin = "TestPlugin")
resourceControl.insertThreshold(siteName = 'site1', taskType = 'Processing', \
maxSlots = 10000)
inputFileset = Fileset("input")
inputFileset.create()
inputFileA = File("lfnA", locations = "goodse.cern.ch")
inputFileB = File("lfnB", locations = "goodse.cern.ch")
inputFileC = File("lfnC", locations = "goodse.cern.ch")
inputFileA.create()
inputFileB.create()
inputFileC.create()
inputFileset.addFile(inputFileA)
inputFileset.addFile(inputFileB)
inputFileset.addFile(inputFileC)
inputFileset.commit()
unmergedOutputFileset = Fileset("unmerged")
unmergedOutputFileset.create()
unmergedFileA = File("ulfnA", locations = "goodse.cern.ch")
unmergedFileB = File("ulfnB", locations = "goodse.cern.ch")
unmergedFileC = File("ulfnC", locations = "goodse.cern.ch")
unmergedFileA.create()
unmergedFileB.create()
unmergedFileC.create()
unmergedOutputFileset.addFile(unmergedFileA)
unmergedOutputFileset.addFile(unmergedFileB)
unmergedOutputFileset.addFile(unmergedFileC)
unmergedOutputFileset.commit()
mainProcWorkflow = Workflow(spec = "spec1", owner = "Steve",
name = "Main", task = "Proc")
mainProcWorkflow.create()
mainProcMergeWorkflow = Workflow(spec = "spec1", owner = "Steve",
name = "Main", task = "ProcMerge")
mainProcMergeWorkflow.create()
mainCleanupWorkflow = Workflow(spec = "spec1", owner = "Steve",
name = "Main", task = "Cleanup")
mainCleanupWorkflow.create()
self.mainProcSub = Subscription(fileset = inputFileset,
workflow = mainProcWorkflow,
type = "Processing")
self.mainProcSub.create()
self.mainProcSub.acquireFiles(inputFileA)
self.mainProcSub.completeFiles(inputFileB)
procJobGroup = JobGroup(subscription = self.mainProcSub)
procJobGroup.create()
self.procJobA = Job(name = "ProcJobA")
self.procJobA["state"] = "new"
self.procJobA["location"] = "site1"
self.procJobB = Job(name = "ProcJobB")
self.procJobB["state"] = "executing"
self.procJobB["location"] = "site1"
self.procJobC = Job(name = "ProcJobC")
self.procJobC["state"] = "complete"
self.procJobC["location"] = "site1"
self.procJobA.create(procJobGroup)
self.procJobB.create(procJobGroup)
self.procJobC.create(procJobGroup)
self.mainMergeSub = Subscription(fileset = unmergedOutputFileset,
workflow = mainProcMergeWorkflow,
type = "Merge")
self.mainMergeSub.create()
self.mainMergeSub.acquireFiles(unmergedFileA)
self.mainMergeSub.failFiles(unmergedFileB)
mergeJobGroup = JobGroup(subscription = self.mainMergeSub)
mergeJobGroup.create()
self.mergeJobA = Job(name = "MergeJobA")
self.mergeJobA["state"] = "exhausted"
self.mergeJobA["location"] = "site1"
self.mergeJobB = Job(name = "MergeJobB")
self.mergeJobB["state"] = "cleanout"
self.mergeJobB["location"] = "site1"
self.mergeJobC = Job(name = "MergeJobC")
self.mergeJobC["state"] = "new"
self.mergeJobC["location"] = "site1"
self.mergeJobA.create(mergeJobGroup)
self.mergeJobB.create(mergeJobGroup)
self.mergeJobC.create(mergeJobGroup)
self.mainCleanupSub = Subscription(fileset = unmergedOutputFileset,
workflow = mainCleanupWorkflow,
type = "Cleanup")
self.mainCleanupSub.create()
self.mainCleanupSub.acquireFiles(unmergedFileA)
self.mainCleanupSub.completeFiles(unmergedFileB)
cleanupJobGroup = JobGroup(subscription = self.mainCleanupSub)
cleanupJobGroup.create()
self.cleanupJobA = Job(name = "CleanupJobA")
self.cleanupJobA["state"] = "new"
self.cleanupJobA["location"] = "site1"
self.cleanupJobB = Job(name = "CleanupJobB")
self.cleanupJobB["state"] = "executing"
self.cleanupJobB["location"] = "site1"
self.cleanupJobC = Job(name = "CleanupJobC")
self.cleanupJobC["state"] = "complete"
self.cleanupJobC["location"] = "site1"
self.cleanupJobA.create(cleanupJobGroup)
self.cleanupJobB.create(cleanupJobGroup)
self.cleanupJobC.create(cleanupJobGroup)
jobList = [self.procJobA, self.procJobB, self.procJobC,
self.mergeJobA, self.mergeJobB, self.mergeJobC,
self.cleanupJobA, self.cleanupJobB, self.cleanupJobC]
changeStateAction.execute(jobList)
if baAPI:
for job in jobList:
job['plugin'] = 'TestPlugin'
job['userdn'] = 'Steve'
job['custom']['location'] = 'site1'
baAPI.createNewJobs(wmbsJobs = jobList)
# We'll create an unrelated workflow to verify that it isn't affected
# by the killing code.
bogusFileset = Fileset("dontkillme")
bogusFileset.create()
bogusFileA = File("bogus/lfnA", locations = "goodse.cern.ch")
bogusFileA.create()
bogusFileset.addFile(bogusFileA)
bogusFileset.commit()
bogusWorkflow = Workflow(spec = "spec2", owner = "Steve",
name = "Bogus", task = "Proc")
bogusWorkflow.create()
self.bogusSub = Subscription(fileset = bogusFileset,
workflow = bogusWorkflow,
type = "Processing")
self.bogusSub.create()
self.bogusSub.acquireFiles(bogusFileA)
return
def verifyFileKillStatus(self):
"""
_verifyFileKillStatus_
Verify that all files were killed correctly. The status of files in
Cleanup and LogCollect subscriptions isn't modified. Status of
already completed and failed files is not modified. Also verify that
the bogus subscription is untouched.
"""
failedFiles = self.mainProcSub.filesOfStatus("Failed")
acquiredFiles = self.mainProcSub.filesOfStatus("Acquired")
completedFiles = self.mainProcSub.filesOfStatus("Completed")
availableFiles = self.mainProcSub.filesOfStatus("Available")
bogusAcquiredFiles = self.bogusSub.filesOfStatus("Acquired")
self.assertEqual(len(availableFiles), 0, \
"Error: There should be no available files.")
self.assertEqual(len(acquiredFiles), 0, \
"Error: There should be no acquired files.")
self.assertEqual(len(bogusAcquiredFiles), 1, \
"Error: There should be one acquired file.")
self.assertEqual(len(completedFiles), 3, \
"Error: There should be only one completed file.")
goldenLFNs = ["lfnA", "lfnB", "lfnC"]
for completedFile in completedFiles:
self.assertTrue(completedFile["lfn"] in goldenLFNs, \
"Error: Extra completed file.")
goldenLFNs.remove(completedFile["lfn"])
self.assertEqual(len(failedFiles), 0, \
"Error: There should be no failed files.")
self.assertEqual(len(goldenLFNs), 0, \
"Error: Missing LFN")
failedFiles = self.mainMergeSub.filesOfStatus("Failed")
acquiredFiles = self.mainMergeSub.filesOfStatus("Acquired")
completedFiles = self.mainMergeSub.filesOfStatus("Completed")
availableFiles = self.mainMergeSub.filesOfStatus("Available")
self.assertEqual(len(acquiredFiles), 0, \
"Error: Merge subscription should have 0 acq files.")
self.assertEqual(len(availableFiles), 0, \
"Error: Merge subscription should have 0 avail files.")
self.assertEqual(len(failedFiles), 1, \
"Error: Merge subscription should have 1 failed files.")
self.assertEqual(list(failedFiles)[0]["lfn"], "ulfnB",
"Error: Wrong failed file.")
self.assertEqual(len(completedFiles), 2, \
"Error: Merge subscription should have 2 compl files.")
goldenLFNs = ["ulfnA", "ulfnC"]
for completedFile in completedFiles:
self.assertTrue(completedFile["lfn"] in goldenLFNs, \
"Error: Extra complete file.")
goldenLFNs.remove(completedFile["lfn"])
self.assertEqual(len(goldenLFNs), 0, \
"Error: Missing LFN")
failedFiles = self.mainCleanupSub.filesOfStatus("Failed")
acquiredFiles = self.mainCleanupSub.filesOfStatus("Acquired")
completedFiles = self.mainCleanupSub.filesOfStatus("Completed")
availableFiles = self.mainCleanupSub.filesOfStatus("Available")
self.assertEqual(len(failedFiles), 0, \
"Error: Cleanup subscription should have 0 fai files.")
self.assertEqual(len(acquiredFiles), 1, \
"Error: There should be only one acquired file.")
self.assertEqual(list(acquiredFiles)[0]["lfn"], "ulfnA", \
"Error: Wrong acquired LFN.")
self.assertEqual(len(completedFiles), 1, \
"Error: There should be only one completed file.")
self.assertEqual(list(completedFiles)[0]["lfn"], "ulfnB", \
"Error: Wrong completed LFN.")
self.assertEqual(len(availableFiles), 1, \
"Error: There should be only one available file.")
self.assertEqual(list(availableFiles)[0]["lfn"], "ulfnC", \
"Error: Wrong completed LFN.")
return
def verifyJobKillStatus(self):
"""
_verifyJobKillStatus_
Verify that jobs are killed correctly. Jobs belonging to Cleanup and
LogCollect subscriptions are not killed. The status of jobs that have
already finished running is not changed.
"""
self.procJobA.load()
self.procJobB.load()
self.procJobC.load()
self.assertEqual(self.procJobA["state"], "killed", \
"Error: Proc job A should be killed.")
self.assertEqual(self.procJobB["state"], "killed", \
"Error: Proc job B should be killed.")
self.assertEqual(self.procJobC["state"], "complete", \
"Error: Proc job C should be complete.")
self.mergeJobA.load()
self.mergeJobB.load()
self.mergeJobC.load()
self.assertEqual(self.mergeJobA["state"], "exhausted", \
"Error: Merge job A should be exhausted.")
self.assertEqual(self.mergeJobB["state"], "cleanout", \
"Error: Merge job B should be cleanout.")
self.assertEqual(self.mergeJobC["state"], "killed", \
"Error: Merge job C should be killed.")
self.cleanupJobA.load()
self.cleanupJobB.load()
self.cleanupJobC.load()
self.assertEqual(self.cleanupJobA["state"], "new", \
"Error: Cleanup job A should be new.")
self.assertEqual(self.cleanupJobB["state"], "executing", \
"Error: Cleanup job B should be executing.")
self.assertEqual(self.cleanupJobC["state"], "complete", \
"Error: Cleanup job C should be complete.")
return
def testKillWorkflow(self):
"""
_testKillWorkflow_
Verify that workflow killing works correctly.
"""
configFile = EmulatorSetup.setupWMAgentConfig()
config = loadConfigurationFile(configFile)
baAPI = BossAirAPI(config = config)
# Create nine jobs
self.setupForKillTest(baAPI = baAPI)
self.assertEqual(len(baAPI._listRunJobs()), 9)
killWorkflow("Main", config, config)
self.verifyFileKillStatus()
self.verifyJobKillStatus()
self.assertEqual(len(baAPI._listRunJobs()), 8)
EmulatorSetup.deleteConfig(configFile)
return
def createTestWMSpec(self):
"""
_createTestWMSpec_
Create a WMSpec that has a processing, merge, cleanup and skims tasks that
can be used by the subscription creation test.
"""
testWorkload = WMWorkloadHelper(WMWorkload("TestWorkload"))
testWorkload.setDashboardActivity("TestReReco")
testWorkload.setSpecUrl("/path/to/workload")
testWorkload.setOwnerDetails("sfoulkes", "DMWM", {'dn': 'MyDN'})
procTask = testWorkload.newTask("ProcessingTask")
procTask.setTaskType("Processing")
procTask.setSplittingAlgorithm("FileBased", files_per_job = 1)
procTaskCMSSW = procTask.makeStep("cmsRun1")
procTaskCMSSW.setStepType("CMSSW")
procTaskCMSSWHelper = procTaskCMSSW.getTypeHelper()
procTask.setTaskType("Processing")
procTask.setSiteWhitelist(["site1"])
procTask.setSiteBlacklist(["site2"])
procTask.applyTemplates()
procTaskCMSSWHelper.addOutputModule("OutputA",
primaryDataset = "bogusPrimary",
processedDataset = "bogusProcessed",
dataTier = "DataTierA",
lfnBase = "bogusUnmerged",
mergedLFNBase = "bogusMerged",
filterName = None)
mergeTask = procTask.addTask("MergeTask")
mergeTask.setInputReference(procTaskCMSSW, outputModule = "OutputA")
mergeTask.setTaskType("Merge")
mergeTask.setSplittingAlgorithm("WMBSMergeBySize", min_merge_size = 1,
max_merge_size = 2, max_merge_events = 3)
mergeTaskCMSSW = mergeTask.makeStep("cmsRun1")
mergeTaskCMSSW.setStepType("CMSSW")
mergeTaskCMSSWHelper = mergeTaskCMSSW.getTypeHelper()
mergeTask.setTaskType("Merge")
mergeTask.applyTemplates()
mergeTaskCMSSWHelper.addOutputModule("Merged",
primaryDataset = "bogusPrimary",
processedDataset = "bogusProcessed",
dataTier = "DataTierA",
lfnBase = "bogusUnmerged",
mergedLFNBase = "bogusMerged",
filterName = None)
cleanupTask = procTask.addTask("CleanupTask")
cleanupTask.setInputReference(procTaskCMSSW, outputModule = "OutputA")
cleanupTask.setTaskType("Merge")
cleanupTask.setSplittingAlgorithm("SiblingProcessingBase", files_per_job = 50)
cleanupTaskCMSSW = cleanupTask.makeStep("cmsRun1")
cleanupTaskCMSSW.setStepType("CMSSW")
cleanupTaskCMSSWHelper = cleanupTaskCMSSW.getTypeHelper()
cleanupTask.setTaskType("Cleanup")
cleanupTask.applyTemplates()
skimTask = mergeTask.addTask("SkimTask")
skimTask.setTaskType("Skim")
skimTask.setInputReference(mergeTaskCMSSW, outputModule = "Merged")
skimTask.setSplittingAlgorithm("FileBased", files_per_job = 1, include_parents = True)
skimTaskCMSSW = skimTask.makeStep("cmsRun1")
skimTaskCMSSW.setStepType("CMSSW")
skimTaskCMSSWHelper = skimTaskCMSSW.getTypeHelper()
skimTask.setTaskType("Skim")
skimTask.applyTemplates()
skimTaskCMSSWHelper.addOutputModule("SkimOutputA",
primaryDataset = "bogusPrimary",
processedDataset = "bogusProcessed",
dataTier = "DataTierA",
lfnBase = "bogusUnmerged",
mergedLFNBase = "bogusMerged",
filterName = None)
skimTaskCMSSWHelper.addOutputModule("SkimOutputB",
primaryDataset = "bogusPrimary",
processedDataset = "bogusProcessed",
dataTier = "DataTierA",
lfnBase = "bogusUnmerged",
mergedLFNBase = "bogusMerged",
filterName = None)
return testWorkload
def testCreateSubscription(self):
"""
_testCreateSubscription_
Verify that the subscription creation code works correctly.
"""
resourceControl = ResourceControl()
resourceControl.insertSite(siteName = 'site1', seName = 'goodse.cern.ch',
ceName = 'site1', plugin = "TestPlugin")
resourceControl.insertSite(siteName = 'site2', seName = 'goodse2.cern.ch',
ceName = 'site2', plugin = "TestPlugin")
testWorkload = self.createTestWMSpec()
testWMBSHelper = WMBSHelper(testWorkload, "SomeBlock", cachepath = self.workDir)
testWMBSHelper.createSubscription()
procWorkflow = Workflow(name = "TestWorkload",
task = "/TestWorkload/ProcessingTask")
procWorkflow.load()
self.assertEqual(procWorkflow.owner, "sfoulkes",
"Error: Wrong owner: %s" % procWorkflow.owner)
self.assertEqual(procWorkflow.group, "DMWM",
"Error: Wrong group: %s" % procWorkflow.group)
self.assertEqual(procWorkflow.wfType, "TestReReco",
"Error: Wrong type.")
self.assertEqual(procWorkflow.spec, os.path.join(self.workDir, procWorkflow.name,
"WMSandbox", "WMWorkload.pkl"),
"Error: Wrong spec URL")
self.assertEqual(len(procWorkflow.outputMap.keys()), 1,
"Error: Wrong number of WF outputs.")
mergedProcOutput = procWorkflow.outputMap["OutputA"][0]["merged_output_fileset"]
unmergedProcOutput = procWorkflow.outputMap["OutputA"][0]["output_fileset"]
mergedProcOutput.loadData()
unmergedProcOutput.loadData()
self.assertEqual(mergedProcOutput.name, "/TestWorkload/ProcessingTask/MergeTask/merged-Merged",
"Error: Merged output fileset is wrong.")
self.assertEqual(unmergedProcOutput.name, "/TestWorkload/ProcessingTask/unmerged-OutputA",
"Error: Unmerged output fileset is wrong.")
mergeWorkflow = Workflow(name = "TestWorkload",
task = "/TestWorkload/ProcessingTask/MergeTask")
mergeWorkflow.load()
self.assertEqual(mergeWorkflow.owner, "sfoulkes",
"Error: Wrong owner.")
self.assertEqual(mergeWorkflow.spec, os.path.join(self.workDir, mergeWorkflow.name,
"WMSandbox", "WMWorkload.pkl"),
"Error: Wrong spec URL")
self.assertEqual(len(mergeWorkflow.outputMap.keys()), 1,
"Error: Wrong number of WF outputs.")
cleanupWorkflow = Workflow(name = "TestWorkload",
task = "/TestWorkload/ProcessingTask/CleanupTask")
cleanupWorkflow.load()
self.assertEqual(cleanupWorkflow.owner, "sfoulkes",
"Error: Wrong owner.")
self.assertEqual(cleanupWorkflow.spec, os.path.join(self.workDir, cleanupWorkflow.name,
"WMSandbox", "WMWorkload.pkl"),
"Error: Wrong spec URL")
self.assertEqual(len(cleanupWorkflow.outputMap.keys()), 0,
"Error: Wrong number of WF outputs.")
unmergedMergeOutput = mergeWorkflow.outputMap["Merged"][0]["output_fileset"]
unmergedMergeOutput.loadData()
self.assertEqual(unmergedMergeOutput.name, "/TestWorkload/ProcessingTask/MergeTask/merged-Merged",
"Error: Unmerged output fileset is wrong.")
skimWorkflow = Workflow(name = "TestWorkload",
task = "/TestWorkload/ProcessingTask/MergeTask/SkimTask")
skimWorkflow.load()
self.assertEqual(skimWorkflow.owner, "sfoulkes",
"Error: Wrong owner.")
self.assertEqual(skimWorkflow.spec, os.path.join(self.workDir, skimWorkflow.name,
"WMSandbox", "WMWorkload.pkl"),
"Error: Wrong spec URL")
self.assertEqual(len(skimWorkflow.outputMap.keys()), 2,
"Error: Wrong number of WF outputs.")
mergedSkimOutputA = skimWorkflow.outputMap["SkimOutputA"][0]["merged_output_fileset"]
unmergedSkimOutputA = skimWorkflow.outputMap["SkimOutputA"][0]["output_fileset"]
mergedSkimOutputB = skimWorkflow.outputMap["SkimOutputB"][0]["merged_output_fileset"]
unmergedSkimOutputB = skimWorkflow.outputMap["SkimOutputB"][0]["output_fileset"]
mergedSkimOutputA.loadData()
mergedSkimOutputB.loadData()
unmergedSkimOutputA.loadData()
unmergedSkimOutputB.loadData()
self.assertEqual(mergedSkimOutputA.name, "/TestWorkload/ProcessingTask/MergeTask/SkimTask/unmerged-SkimOutputA",
"Error: Merged output fileset is wrong: %s" % mergedSkimOutputA.name)
self.assertEqual(unmergedSkimOutputA.name, "/TestWorkload/ProcessingTask/MergeTask/SkimTask/unmerged-SkimOutputA",
"Error: Unmerged output fileset is wrong.")
self.assertEqual(mergedSkimOutputB.name, "/TestWorkload/ProcessingTask/MergeTask/SkimTask/unmerged-SkimOutputB",
"Error: Merged output fileset is wrong.")
self.assertEqual(unmergedSkimOutputB.name, "/TestWorkload/ProcessingTask/MergeTask/SkimTask/unmerged-SkimOutputB",
"Error: Unmerged output fileset is wrong.")
topLevelFileset = Fileset(name = "TestWorkload-ProcessingTask-SomeBlock")
topLevelFileset.loadData()
procSubscription = Subscription(fileset = topLevelFileset, workflow = procWorkflow)
procSubscription.loadData()
self.assertEqual(len(procSubscription.getWhiteBlackList()), 2,
"Error: Wrong site white/black list for proc sub.")
for site in procSubscription.getWhiteBlackList():
if site["site_name"] == "site1":
self.assertEqual(site["valid"], 1,
"Error: Site should be white listed.")
else:
self.assertEqual(site["valid"], 0,
"Error: Site should be black listed.")
self.assertEqual(procSubscription["type"], "Processing",
"Error: Wrong subscription type.")
self.assertEqual(procSubscription["split_algo"], "FileBased",
"Error: Wrong split algo.")
mergeSubscription = Subscription(fileset = unmergedProcOutput, workflow = mergeWorkflow)
mergeSubscription.loadData()
self.assertEqual(len(mergeSubscription.getWhiteBlackList()), 0,
"Error: Wrong white/black list for merge sub.")
self.assertEqual(mergeSubscription["type"], "Merge",
"Error: Wrong subscription type.")
self.assertEqual(mergeSubscription["split_algo"], "WMBSMergeBySize",
"Error: Wrong split algo.")
skimSubscription = Subscription(fileset = unmergedMergeOutput, workflow = skimWorkflow)
skimSubscription.loadData()
self.assertEqual(skimSubscription["type"], "Skim",
"Error: Wrong subscription type.")
self.assertEqual(skimSubscription["split_algo"], "FileBased",
"Error: Wrong split algo.")
return
def testTruncatedWFInsertion(self):
"""
_testTruncatedWFInsertion_
"""
resourceControl = ResourceControl()
resourceControl.insertSite(siteName = 'site1', seName = 'goodse.cern.ch',
ceName = 'site1', plugin = "TestPlugin")
resourceControl.insertSite(siteName = 'site2', seName = 'goodse2.cern.ch',
ceName = 'site2', plugin = "TestPlugin")
testWorkload = self.createTestWMSpec()
testWMBSHelper = WMBSHelper(testWorkload, "SomeBlock", cachepath = self.workDir)
testWMBSHelper.createSubscription()
testWorkload.truncate("ResubmitTestWorkload", "/TestWorkload/ProcessingTask/MergeTask",
"someserver", "somedatabase")
testResubmitWMBSHelper = WMBSHelper(testWorkload, "SomeBlock2", cachepath = self.workDir)
testResubmitWMBSHelper.createSubscription()
mergeWorkflow = Workflow(name = "ResubmitTestWorkload",
task = "/ResubmitTestWorkload/MergeTask")
mergeWorkflow.load()
self.assertEqual(mergeWorkflow.owner, "sfoulkes",
"Error: Wrong owner.")
self.assertEqual(mergeWorkflow.spec, os.path.join(self.workDir, mergeWorkflow.name,
"WMSandbox", "WMWorkload.pkl"),
"Error: Wrong spec URL")
self.assertEqual(len(mergeWorkflow.outputMap.keys()), 1,
"Error: Wrong number of WF outputs.")
cleanupWorkflow = Workflow(name = "ResubmitTestWorkload",
task = "/ResubmitTestWorkload/CleanupTask")
cleanupWorkflow.load()
self.assertEqual(cleanupWorkflow.owner, "sfoulkes",
"Error: Wrong owner.")
self.assertEqual(cleanupWorkflow.spec, os.path.join(self.workDir, cleanupWorkflow.name,
"WMSandbox", "WMWorkload.pkl"),
"Error: Wrong spec URL")
self.assertEqual(len(cleanupWorkflow.outputMap.keys()), 0,
"Error: Wrong number of WF outputs.")
unmergedMergeOutput = mergeWorkflow.outputMap["Merged"][0]["output_fileset"]
unmergedMergeOutput.loadData()
self.assertEqual(unmergedMergeOutput.name, "/ResubmitTestWorkload/MergeTask/merged-Merged",
"Error: Unmerged output fileset is wrong.")
skimWorkflow = Workflow(name = "ResubmitTestWorkload",
task = "/ResubmitTestWorkload/MergeTask/SkimTask")
skimWorkflow.load()
self.assertEqual(skimWorkflow.owner, "sfoulkes",
"Error: Wrong owner.")
self.assertEqual(skimWorkflow.spec, os.path.join(self.workDir, skimWorkflow.name,
"WMSandbox", "WMWorkload.pkl"),
"Error: Wrong spec URL")
self.assertEqual(len(skimWorkflow.outputMap.keys()), 2,
"Error: Wrong number of WF outputs.")
mergedSkimOutputA = skimWorkflow.outputMap["SkimOutputA"][0]["merged_output_fileset"]
unmergedSkimOutputA = skimWorkflow.outputMap["SkimOutputA"][0]["output_fileset"]
mergedSkimOutputB = skimWorkflow.outputMap["SkimOutputB"][0]["merged_output_fileset"]
unmergedSkimOutputB = skimWorkflow.outputMap["SkimOutputB"][0]["output_fileset"]
mergedSkimOutputA.loadData()
mergedSkimOutputB.loadData()
unmergedSkimOutputA.loadData()
unmergedSkimOutputB.loadData()
self.assertEqual(mergedSkimOutputA.name, "/ResubmitTestWorkload/MergeTask/SkimTask/unmerged-SkimOutputA",
"Error: Merged output fileset is wrong: %s" % mergedSkimOutputA.name)
self.assertEqual(unmergedSkimOutputA.name, "/ResubmitTestWorkload/MergeTask/SkimTask/unmerged-SkimOutputA",
"Error: Unmerged output fileset is wrong.")
self.assertEqual(mergedSkimOutputB.name, "/ResubmitTestWorkload/MergeTask/SkimTask/unmerged-SkimOutputB",
"Error: Merged output fileset is wrong.")
self.assertEqual(unmergedSkimOutputB.name, "/ResubmitTestWorkload/MergeTask/SkimTask/unmerged-SkimOutputB",
"Error: Unmerged output fileset is wrong.")
topLevelFileset = Fileset(name = "ResubmitTestWorkload-MergeTask-SomeBlock2")
topLevelFileset.loadData()
mergeSubscription = Subscription(fileset = topLevelFileset, workflow = mergeWorkflow)
mergeSubscription.loadData()
self.assertEqual(len(mergeSubscription.getWhiteBlackList()), 0,
"Error: Wrong white/black list for merge sub.")
self.assertEqual(mergeSubscription["type"], "Merge",
"Error: Wrong subscription type.")
self.assertEqual(mergeSubscription["split_algo"], "WMBSMergeBySize",
"Error: Wrong split algo.")
skimSubscription = Subscription(fileset = unmergedMergeOutput, workflow = skimWorkflow)
skimSubscription.loadData()
self.assertEqual(skimSubscription["type"], "Skim",
"Error: Wrong subscription type.")
self.assertEqual(skimSubscription["split_algo"], "FileBased",
"Error: Wrong split algo.")
return
def setupMCWMSpec(self):
"""Setup MC workflow"""
self.wmspec = self.createMCWMSpec()
self.topLevelTask = getFirstTask(self.wmspec)
self.inputDataset = self.topLevelTask.inputDataset()
self.dataset = self.topLevelTask.getInputDatasetPath()
self.dbs = None
self.siteDB = fakeSiteDB()
def createWMSpec(self, name = 'ReRecoWorkload'):
wmspec = rerecoWorkload(name, rerecoArgs)
wmspec.setSpecUrl("/path/to/workload")
return wmspec
def createMCWMSpec(self, name = 'MonteCarloWorkload'):
wmspec = monteCarloWorkload(name, mcArgs)
wmspec.setSpecUrl("/path/to/workload")
getFirstTask(wmspec).addProduction(totalevents = 10000)
return wmspec
def getDBS(self, wmspec):
topLevelTask = getFirstTask(wmspec)
inputDataset = topLevelTask.inputDataset()
dbs = MockDBSReader(inputDataset.dbsurl)
#dbsDict = {self.inputDataset.dbsurl : self.dbs}
return dbs
def createWMBSHelperWithTopTask(self, wmspec, block, mask = None):
topLevelTask = getFirstTask(wmspec)
wmbs = WMBSHelper(wmspec, block, mask, cachepath = self.workDir)
if block:
block = self.dbs.getFileBlock(block)[block]
wmbs.createSubscriptionAndAddFiles(block = block)
return wmbs
# def testProduction(self):
# """Production workflow"""
# pass
def testReReco(self):
"""ReReco workflow"""
# create workflow
block = self.dataset + "#1"
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
files = wmbs.validFiles(self.dbs.getFileBlock(block))
self.assertEqual(len(files), 1)
def testReRecoBlackRunRestriction(self):
"""ReReco workflow with Run restrictions"""
block = self.dataset + "#2"
#add run blacklist
self.topLevelTask.setInputRunBlacklist([1, 2, 3, 4])
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
files = wmbs.validFiles(self.dbs.getFileBlock(block)[block]['Files'])
self.assertEqual(len(files), 0)
def testReRecoWhiteRunRestriction(self):
block = self.dataset + "#2"
# Run Whitelist
self.topLevelTask.setInputRunWhitelist([2])
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
files = wmbs.validFiles(self.dbs.getFileBlock(block)[block]['Files'])
self.assertEqual(len(files), GlobalParams.numOfFilesPerBlock())
def testDuplicateFileInsert(self):
# using default wmspec
block = self.dataset + "#1"
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, block)
wmbs.topLevelFileset.loadData()
numOfFiles = len(wmbs.topLevelFileset.files)
# check initially inserted files.
dbsFiles = self.dbs.getFileBlock(block)[block]['Files']
self.assertEqual(numOfFiles, len(dbsFiles))
firstFileset = wmbs.topLevelFileset
wmbsDao = wmbs.daofactory(classname = "Files.InFileset")
numOfFiles = len(wmbsDao.execute(firstFileset.id))
self.assertEqual(numOfFiles, len(dbsFiles))
# use the new spec with same inputdataset
block = self.dataset + "#1"
wmspec = self.createWMSpec("TestSpec1")
dbs = self.getDBS(wmspec)
wmbs = self.createWMBSHelperWithTopTask(wmspec, block)
# check duplicate insert
dbsFiles = dbs.getFileBlock(block)[block]['Files']
numOfFiles = wmbs.addFiles(dbs.getFileBlock(block)[block])
self.assertEqual(numOfFiles, 0)
secondFileset = wmbs.topLevelFileset
wmbsDao = wmbs.daofactory(classname = "Files.InFileset")
numOfFiles = len(wmbsDao.execute(secondFileset.id))
self.assertEqual(numOfFiles, len(dbsFiles))
self.assertNotEqual(firstFileset.id, secondFileset.id)
def testParentage(self):
"""
TODO: add the parentage test.
1. check whether parent files are created in wmbs.
2. check parent files are associated to child.
3. When 2 specs with the same input data (one with parent processing, one without it)
is inserted, if one without parent processing inserted first then the other with
parent processing insert, it still needs to create parent files although child files
are duplicate
"""
pass
def testMCFakeFileInjection(self):
"""Inject fake Monte Carlo files into WMBS"""
self.setupMCWMSpec()
mask = Mask(FirstRun = 12, FirstLumi = 1234, FirstEvent = 12345,
LastEvent = 999995, LastLumi = 12345, LastRun = 12)
# add sites that would normally be added by operator via resource_control
locationDAO = self.daoFactory(classname = "Locations.New")
ses = []
for site in ['T2_XX_SiteA', 'T2_XX_SiteB']:
locationDAO.execute(siteName = site, seName = self.siteDB.cmsNametoSE(site))
ses.append(self.siteDB.cmsNametoSE(site))
wmbs = self.createWMBSHelperWithTopTask(self.wmspec, None, mask)
subscription = wmbs.topLevelSubscription
self.assertEqual(1, subscription.exists())
fileset = subscription['fileset']
self.assertEqual(1, fileset.exists())
fileset.loadData() # need to refresh from database
self.assertEqual(len(fileset.files), 1)
self.assertEqual(len(fileset.parents), 0)
self.assertFalse(fileset.open)
file = list(fileset.files)[0]
self.assertEqual(file['events'], mask['LastEvent'] - mask['FirstEvent'] + 1) # inclusive range
self.assertEqual(file['merged'], False) # merged files get added to dbs
self.assertEqual(len(file['parents']), 0)
#file.loadData()
self.assertEqual(sorted(file['locations']), sorted(ses))
self.assertEqual(len(file.getParentLFNs()), 0)
self.assertEqual(len(file.getRuns()), 1)
run = file.getRuns()[0]
self.assertEqual(run.run, mask['FirstRun'])
self.assertEqual(run.lumis[0], mask['FirstLumi'])
self.assertEqual(run.lumis[-1], mask['LastLumi'])
self.assertEqual(len(run.lumis), mask['LastLumi'] - mask['FirstLumi'] + 1)
class ReportIntegrationTest(unittest.TestCase):
"""
_ReportIntegrationTest_
"""
def setUp(self):
"""
_setUp_
Setup the database and WMBS for the test.
"""
self.testInit = TestInit(__file__)
self.testInit.setLogging()
self.testInit.setDatabaseConnection()
self.testInit.setSchema(customModules = ["WMComponent.DBS3Buffer",
"WMCore.WMBS"],
useDefault = False)
myThread = threading.currentThread()
self.daofactory = DAOFactory(package = "WMCore.WMBS",
logger = myThread.logger,
dbinterface = myThread.dbi)
self.dbsfactory = DAOFactory(package = "WMComponent.DBS3Buffer",
logger = myThread.logger,
dbinterface = myThread.dbi)
locationAction = self.daofactory(classname = "Locations.New")
locationAction.execute(siteName = "site1", pnn = "T1_US_FNAL_Disk")
inputFile = File(lfn = "/path/to/some/lfn", size = 10, events = 10,
locations = "T1_US_FNAL_Disk")
inputFile.create()
inputFileset = Fileset(name = "InputFileset")
inputFileset.create()
inputFileset.addFile(inputFile)
inputFileset.commit()
unmergedFileset = Fileset(name = "UnmergedFileset")
unmergedFileset.create()
mergedFileset = Fileset(name = "MergedFileset")
mergedFileset.create()
procWorkflow = Workflow(spec = "wf001.xml", owner = "Steve",
name = "TestWF", task = "/TestWF/None")
procWorkflow.create()
procWorkflow.addOutput("outputRECORECO", unmergedFileset)
mergeWorkflow = Workflow(spec = "wf002.xml", owner = "Steve",
name = "MergeWF", task = "/MergeWF/None")
mergeWorkflow.create()
mergeWorkflow.addOutput("Merged", mergedFileset)
insertWorkflow = self.dbsfactory(classname = "InsertWorkflow")
insertWorkflow.execute("TestWF", "/TestWF/None", 0, 0, 0, 0)
insertWorkflow.execute("MergeWF", "/MergeWF/None", 0, 0, 0, 0)
self.procSubscription = Subscription(fileset = inputFileset,
workflow = procWorkflow,
split_algo = "FileBased",
type = "Processing")
self.procSubscription.create()
self.procSubscription.acquireFiles()
self.mergeSubscription = Subscription(fileset = unmergedFileset,
workflow = mergeWorkflow,
split_algo = "WMBSMergeBySize",
type = "Merge")
self.mergeSubscription.create()
self.procJobGroup = JobGroup(subscription = self.procSubscription)
self.procJobGroup.create()
self.mergeJobGroup = JobGroup(subscription = self.mergeSubscription)
self.mergeJobGroup.create()
self.testJob = Job(name = "testJob", files = [inputFile])
self.testJob.create(group = self.procJobGroup)
self.testJob["state"] = "complete"
myThread = threading.currentThread()
self.daofactory = DAOFactory(package = "WMCore.WMBS",
logger = myThread.logger,
dbinterface = myThread.dbi)
self.stateChangeAction = self.daofactory(classname = "Jobs.ChangeState")
self.setFWJRAction = self.daofactory(classname = "Jobs.SetFWJRPath")
self.getJobTypeAction = self.daofactory(classname = "Jobs.GetType")
locationAction = self.daofactory(classname = "Locations.New")
locationAction.execute(siteName = "cmssrm.fnal.gov")
self.stateChangeAction.execute(jobs = [self.testJob])
self.tempDir = tempfile.mkdtemp()
return
def tearDown(self):
"""
_tearDown_
Clear out the database and the pickled report file.
"""
self.testInit.clearDatabase()
try:
os.remove(os.path.join(self.tempDir, "ProcReport.pkl"))
os.remove(os.path.join(self.tempDir, "MergeReport.pkl"))
except Exception as ex:
pass
try:
os.rmdir(self.tempDir)
except Exception as ex:
pass
return
def createConfig(self, workerThreads):
"""
_createConfig_
Create a config for the JobAccountant with the given number of worker
threads. This config needs to include information for connecting to the
database as the component will create it's own database connections.
These parameters are still pulled from the environment.
"""
config = self.testInit.getConfiguration()
self.testInit.generateWorkDir(config)
config.section_("JobStateMachine")
config.JobStateMachine.couchurl = os.getenv("COUCHURL")
config.JobStateMachine.couchDBName = "report_integration_t"
config.JobStateMachine.jobSummaryDBName = "report_integration_wmagent_summary_t"
config.component_("JobAccountant")
config.JobAccountant.pollInterval = 60
config.JobAccountant.workerThreads = workerThreads
config.JobAccountant.componentDir = os.getcwd()
config.JobAccountant.logLevel = 'SQLDEBUG'
config.component_("TaskArchiver")
config.TaskArchiver.localWMStatsURL = "%s/%s" % (config.JobStateMachine.couchurl, config.JobStateMachine.jobSummaryDBName)
return config
def verifyJobSuccess(self, jobID):
"""
_verifyJobSuccess_
Verify that the metadata for a successful job is correct. This will
check the outcome, retry count and state.
"""
testJob = Job(id = jobID)
testJob.load()
assert testJob["state"] == "success", \
"Error: test job in wrong state: %s" % testJob["state"]
assert testJob["retry_count"] == 0, \
"Error: test job has wrong retry count: %s" % testJob["retry_count"]
assert testJob["outcome"] == "success", \
"Error: test job has wrong outcome: %s" % testJob["outcome"]
return
def verifyFileMetaData(self, jobID, fwkJobReportFiles):
"""
_verifyFileMetaData_
Verify that all the files that were output by a job made it into WMBS
correctly. Compare the contents of WMBS to the files in the frameworks
job report.
Note that fwkJobReportFiles is a list of DataStructs File objects.
"""
testJob = Job(id = jobID)
testJob.loadData()
inputLFNs = []
for inputFile in testJob["input_files"]:
inputLFNs.append(inputFile["lfn"])
for fwkJobReportFile in fwkJobReportFiles:
outputFile = File(lfn = fwkJobReportFile["lfn"])
outputFile.loadData(parentage = 1)
assert outputFile["events"] == int(fwkJobReportFile["events"]), \
"Error: Output file has wrong events: %s, %s" % \
(outputFile["events"], fwkJobReportFile["events"])
assert outputFile["size"] == int(fwkJobReportFile["size"]), \
"Error: Output file has wrong size: %s, %s" % \
(outputFile["size"], fwkJobReportFile["size"])
for ckType in fwkJobReportFile["checksums"]:
assert ckType in outputFile["checksums"], \
"Error: Output file is missing checksums: %s" % ckType
assert outputFile["checksums"][ckType] == fwkJobReportFile["checksums"][ckType], \
"Error: Checksums don't match."
assert len(fwkJobReportFile["checksums"]) == \
len(outputFile["checksums"]), \
"Error: Wrong number of checksums."
jobType = self.getJobTypeAction.execute(jobID = jobID)
if jobType == "Merge":
assert str(outputFile["merged"]) == "True", \
"Error: Merge jobs should output merged files."
else:
assert outputFile["merged"] == fwkJobReportFile["merged"], \
"Error: Output file merged output is wrong: %s, %s" % \
(outputFile["merged"], fwkJobReportFile["merged"])
assert len(outputFile["locations"]) == 1, \
"Error: outputfile should have one location: %s" % outputFile["locations"]
assert list(outputFile["locations"])[0] == list(fwkJobReportFile["locations"])[0], \
"Error: wrong location for file."
assert len(outputFile["parents"]) == len(inputLFNs), \
"Error: Output file has wrong number of parents."
for outputParent in outputFile["parents"]:
assert outputParent["lfn"] in inputLFNs, \
"Error: Unknown parent file: %s" % outputParent["lfn"]
fwjrRuns = {}
for run in fwkJobReportFile["runs"]:
fwjrRuns[run.run] = run.lumis
for run in outputFile["runs"]:
assert run.run in fwjrRuns, \
"Error: Extra run in output: %s" % run.run
for lumi in run:
assert lumi in fwjrRuns[run.run], \
"Error: Extra lumi: %s" % lumi
fwjrRuns[run.run].remove(lumi)
if len(fwjrRuns[run.run]) == 0:
del fwjrRuns[run.run]
assert len(fwjrRuns) == 0, \
"Error: Missing runs, lumis: %s" % fwjrRuns
testJobGroup = JobGroup(id = testJob["jobgroup"])
testJobGroup.loadData()
jobGroupFileset = testJobGroup.output
jobGroupFileset.loadData()
assert outputFile["id"] in jobGroupFileset.getFiles(type = "id"), \
"Error: output file not in jobgroup fileset."
if testJob["mask"]["FirstEvent"] == None:
assert outputFile["first_event"] == 0, \
"Error: first event not set correctly: 0, %s" % \
outputFile["first_event"]
else:
assert testJob["mask"]["FirstEvent"] == outputFile["first_event"], \
"Error: last event not set correctly: %s, %s" % \
(testJob["mask"]["FirstEvent"], outputFile["first_event"])
return
def testReportHandling(self):
"""
_testReportHandling_
Verify that we're able to parse a CMSSW report, convert it to a Report()
style report, pickle it and then have the accountant process it.
"""
self.procPath = os.path.join(WMCore.WMBase.getTestBase(),
"WMCore_t/FwkJobReport_t/CMSSWProcessingReport.xml")
myReport = Report("cmsRun1")
myReport.parse(self.procPath)
# Fake some metadata that should be added by the stageout scripts.
for fileRef in myReport.getAllFileRefsFromStep("cmsRun1"):
fileRef.size = 1024
fileRef.location = "cmssrm.fnal.gov"
fwjrPath = os.path.join(self.tempDir, "ProcReport.pkl")
cmsRunStep = myReport.retrieveStep("cmsRun1")
cmsRunStep.status = 0
myReport.setTaskName('/TestWF/None')
myReport.persist(fwjrPath)
self.setFWJRAction.execute(jobID = self.testJob["id"], fwjrPath = fwjrPath)
pFile = DBSBufferFile(lfn = "/path/to/some/lfn", size = 600000, events = 60000)
pFile.setAlgorithm(appName = "cmsRun", appVer = "UNKNOWN",
appFam = "RECO", psetHash = "GIBBERISH",
configContent = "MOREGIBBERISH")
pFile.setDatasetPath("/bogus/dataset/path")
#pFile.addRun(Run(1, *[45]))
pFile.create()
config = self.createConfig(workerThreads = 1)
accountant = JobAccountantPoller(config)
accountant.setup()
accountant.algorithm()
self.verifyJobSuccess(self.testJob["id"])
self.verifyFileMetaData(self.testJob["id"], myReport.getAllFilesFromStep("cmsRun1"))
inputFile = File(lfn = "/store/backfill/2/unmerged/WMAgentCommissioining10/MinimumBias/RECO/rereco_GR09_R_34X_V5_All_v1/0000/outputRECORECO.root")
inputFile.load()
self.testMergeJob = Job(name = "testMergeJob", files = [inputFile])
self.testMergeJob.create(group = self.mergeJobGroup)
self.testMergeJob["state"] = "complete"
self.stateChangeAction.execute(jobs = [self.testMergeJob])
self.mergePath = os.path.join(WMCore.WMBase.getTestBase(),
"WMCore_t/FwkJobReport_t/CMSSWMergeReport.xml")
myReport = Report("mergeReco")
myReport.parse(self.mergePath)
# Fake some metadata that should be added by the stageout scripts.
for fileRef in myReport.getAllFileRefsFromStep("mergeReco"):
fileRef.size = 1024
fileRef.location = "cmssrm.fnal.gov"
fileRef.dataset = {"applicationName": "cmsRun", "applicationVersion": "CMSSW_3_4_2_patch1",
"primaryDataset": "MinimumBias", "processedDataset": "Rereco-v1",
"dataTier": "RECO"}
fwjrPath = os.path.join(self.tempDir, "MergeReport.pkl")
myReport.setTaskName('/MergeWF/None')
cmsRunStep = myReport.retrieveStep("mergeReco")
cmsRunStep.status = 0
myReport.persist(fwjrPath)
self.setFWJRAction.execute(jobID = self.testMergeJob["id"], fwjrPath = fwjrPath)
accountant.algorithm()
self.verifyJobSuccess(self.testMergeJob["id"])
self.verifyFileMetaData(self.testMergeJob["id"], myReport.getAllFilesFromStep("mergeReco"))
return
