def loadData(self):
"""
_loadData_
Load all information about the job, including the mask and all input
files. Either the ID or the name must be specified before this is
called.
"""
existingTransaction = self.beginTransaction()
self.load()
self.getMask()
fileAction = self.daofactory(classname="Jobs.LoadFiles")
files = fileAction.execute(self["id"],
conn=self.getDBConn(),
transaction=self.existingTransaction())
self["input_files"] = []
for file in files:
newFile = File(id=file["id"])
newFile.loadData(parentage=0)
self.addFile(newFile)
self.commitTransaction(existingTransaction)
return
def testSetLocationByLFN(self):
"""
_testSetLocationByLFN_
Create a file and add a couple locations. Load the file from the
database to make sure that the locations were set correctly.
"""
testFileA = File(lfn = "/this/is/a/lfnA", size = 1024, events = 10,
checksums = {'cksum':1})
testFileA.addRun(Run( 1, *[45]))
testFileA.create()
testFileB = File(lfn = "/this/is/a/lfnB", size = 1024, events = 10,
checksums = {'cksum':1})
testFileB.addRun(Run( 1, *[45]))
testFileB.create()
parentAction = self.daofactory(classname = "Files.SetLocationByLFN")
binds = [{'lfn': "/this/is/a/lfnA", 'location': 'se1.fnal.gov'},
{'lfn': "/this/is/a/lfnB", 'location': 'se1.fnal.gov'}]
parentAction.execute(lfn = binds)
testFileC = File(id = testFileA["id"])
testFileC.loadData()
testFileD = File(id = testFileB["id"])
testFileD.loadData()
self.assertEqual(testFileC['locations'], set(['se1.fnal.gov']))
self.assertEqual(testFileD['locations'], set(['se1.fnal.gov']))
return
def testSetLocation(self):
"""
_testSetLocation_
Create a file and add a couple locations. Load the file from the
database to make sure that the locations were set correctly.
"""
testFileA = File(lfn = "/this/is/a/lfn", size = 1024, events = 10,
checksums = {'cksum':1})
testFileA.addRun(Run( 1, *[45]))
testFileA.create()
testFileA.setLocation(["se1.fnal.gov", "se1.cern.ch"])
testFileA.setLocation(["bunkse1.fnal.gov", "bunkse1.cern.ch"],
immediateSave = False)
testFileB = File(id = testFileA["id"])
testFileB.loadData()
goldenLocations = ["se1.fnal.gov", "se1.cern.ch"]
for location in testFileB["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
return
def testCreateWithLocation(self):
"""
_testCreateWithLocation_
Create a file and add a couple locations. Load the file from the
database to make sure that the locations were set correctly.
"""
testFileA = File(lfn = "/this/is/a/lfn", size = 1024, events = 10,
checksums = {'cksum':1},
locations = set(["T1_US_FNAL_Disk", "T2_CH_CERN"]))
testFileA.addRun(Run( 1, *[45]))
testFileA.create()
testFileB = File(id = testFileA["id"])
testFileB.loadData()
goldenLocations = ["T1_US_FNAL_Disk", "T2_CH_CERN"]
for location in testFileB["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
return
def loadData(self, parentage=1):
"""
_loadData_
Load all the files that belong to this fileset.
"""
existingTransaction = self.beginTransaction()
if self.name is None or self.id < 0:
self.load()
action = self.daofactory(classname="Files.InFileset")
results = action.execute(fileset=self.id,
conn=self.getDBConn(),
transaction=self.existingTransaction())
self.files = set()
self.newfiles = set()
for result in results:
thisFile = File(id=result["fileid"])
thisFile.loadData(parentage=parentage)
self.files.add(thisFile)
self.commitTransaction(existingTransaction)
return
def testSetLocation(self):
"""
_testSetLocation_
Create a file and add a couple locations. Load the file from the
database to make sure that the locations were set correctly.
"""
testFileA = File(lfn="/this/is/a/lfn",
size=1024,
events=10,
checksums={'cksum': 1})
testFileA.addRun(Run(1, *[45]))
testFileA.create()
testFileA.setLocation(["se1.fnal.gov", "se1.cern.ch"])
testFileA.setLocation(["bunkse1.fnal.gov", "bunkse1.cern.ch"],
immediateSave=False)
testFileB = File(id=testFileA["id"])
testFileB.loadData()
goldenLocations = ["se1.fnal.gov", "se1.cern.ch"]
for location in testFileB["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
return
def testAddChild(self):
"""
_testAddChild_
Add a child to some parent files and make sure that all the parentage
information is loaded/stored correctly from the database.
"""
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()
testFileParentA.addChild("/this/is/a/lfn")
testFileParentB.addChild("/this/is/a/lfn")
testFileB = File(id=testFileA["id"])
testFileB.loadData(parentage=1)
goldenFiles = [testFileParentA, testFileParentB]
for parentFile in testFileB["parents"]:
assert parentFile in goldenFiles, "ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, "ERROR: Some parents are missing"
return
def loadData(self):
"""
_loadData_
Load all the files that belong to this fileset.
"""
existingTransaction = self.beginTransaction()
if self.name == None or self.id < 0:
self.load()
action = self.daofactory(classname = "Files.InFileset")
results = action.execute(fileset = self.id,
conn = self.getDBConn(),
transaction = self.existingTransaction())
self.files = set()
self.newfiles = set()
for result in results:
file = File(id = result["fileid"])
file.loadData(parentage = 1)
self.files.add(file)
self.commitTransaction(existingTransaction)
return
def loadData(self):
"""
_loadData_
Load all information about the job, including the mask and all input
files. Either the ID or the name must be specified before this is
called.
"""
existingTransaction = self.beginTransaction()
self.load()
self.getMask()
fileAction = self.daofactory(classname = "Jobs.LoadFiles")
files = fileAction.execute(self["id"], conn = self.getDBConn(),
transaction = self.existingTransaction())
self["input_files"] = []
for file in files:
newFile = File(id = file["id"])
newFile.loadData(parentage = 0)
self.addFile(newFile)
self.commitTransaction(existingTransaction)
return
def testBulkParentage(self):
"""
_testBulkParentage_
Verify that the bulk parentage dao correctly sets file parentage.
"""
testFileChildA = File(lfn = "/this/is/a/child/lfnA", size = 1024,
events = 20, checksums = {'cksum': 1})
testFileChildB = File(lfn = "/this/is/a/child/lfnB", size = 1024,
events = 20, checksums = {'cksum': 1})
testFileChildA.create()
testFileChildB.create()
testFileA = File(lfn = "/this/is/a/lfnA", size = 1024, events = 10,
checksums = {'cksum':1})
testFileA.create()
testFileB = File(lfn = "/this/is/a/lfnB", size = 1024, events = 10,
checksums = {'cksum':1})
testFileB.create()
testFileC = File(lfn = "/this/is/a/lfnC", size = 1024, events = 10,
checksums = {'cksum':1})
testFileC.create()
parentage = [{"child": testFileChildA["id"], "parent": testFileA["id"]},
{"child": testFileChildA["id"], "parent": testFileB["id"]},
{"child": testFileChildA["id"], "parent": testFileC["id"]},
{"child": testFileChildB["id"], "parent": testFileA["id"]},
{"child": testFileChildB["id"], "parent": testFileB["id"]}]
myThread = threading.currentThread()
daofactory = DAOFactory(package = "WMCore.WMBS",
logger = myThread.logger,
dbinterface = myThread.dbi)
bulkParentageAction = daofactory(classname = "Files.AddBulkParentage")
bulkParentageAction.execute(parentage)
testFileD = File(id = testFileChildA["id"])
testFileD.loadData(parentage = 1)
testFileE = File(id = testFileChildB["id"])
testFileE.loadData(parentage = 1)
goldenFiles = [testFileA, testFileB, testFileC]
for parentFile in testFileD["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
goldenFiles = [testFileA, testFileB]
for parentFile in testFileE["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
return
def testAddChildTransaction(self):
"""
_testAddChildTransaction_
Add a child to some parent files and make sure that all the parentage
information is loaded/stored correctly from the database. Rollback the
addition of one of the childs and then verify that it does in fact only
have one parent.
"""
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()
testFileParentA.addChild("/this/is/a/lfn")
myThread = threading.currentThread()
myThread.transaction.begin()
testFileParentB.addChild("/this/is/a/lfn")
testFileB = File(id = testFileA["id"])
testFileB.loadData(parentage = 1)
goldenFiles = [testFileParentA, testFileParentB]
for parentFile in testFileB["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
myThread.transaction.rollback()
testFileB.loadData(parentage = 1)
goldenFiles = [testFileParentA]
for parentFile in testFileB["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
return
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 testSetLocationTransaction(self):
"""
_testSetLocationTransaction_
Create a file at specific locations and commit everything to the
database. Reload the file from the database and verify that the
locations are correct. Rollback the database transaction and once
again reload the file. Verify that the original locations are back.
"""
testFileA = File(lfn="/this/is/a/lfn",
size=1024,
events=10,
checksums={'cksum': 1})
testFileA.addRun(Run(1, *[45]))
testFileA.create()
testFileA.setLocation(["se1.fnal.gov"])
myThread = threading.currentThread()
myThread.transaction.begin()
testFileA.setLocation(["se1.cern.ch"])
testFileA.setLocation(["bunkse1.fnal.gov", "bunkse1.cern.ch"],
immediateSave=False)
testFileB = File(id=testFileA["id"])
testFileB.loadData()
goldenLocations = ["se1.fnal.gov", "se1.cern.ch"]
for location in testFileB["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
myThread.transaction.rollback()
testFileB.loadData()
goldenLocations = ["se1.fnal.gov"]
for location in testFileB["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
return
def testLocationsConstructor(self):
"""
_testLocationsConstructor_
Test to make sure that locations passed into the File() constructor
are loaded from and save to the database correctly. Also test to make
sure that the class behaves well when the location is passed in as a
single string instead of a set.
"""
testFileA = File(lfn="/this/is/a/lfn",
size=1024,
events=10,
checksums={'cksum': 1},
locations=set(["se1.fnal.gov"]))
testFileA.addRun(Run(1, *[45]))
testFileA.create()
testFileB = File(lfn="/this/is/a/lfn2",
size=1024,
events=10,
checksums={'cksum': 1},
locations="se1.fnal.gov")
testFileB.addRun(Run(1, *[45]))
testFileB.create()
testFileC = File(id=testFileA["id"])
testFileC.loadData()
goldenLocations = ["se1.fnal.gov"]
for location in testFileC["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
testFileC = File(id=testFileB["id"])
testFileC.loadData()
goldenLocations = ["se1.fnal.gov"]
for location in testFileC["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
return
def testSetLocationTransaction(self):
"""
_testSetLocationTransaction_
Create a file at specific locations and commit everything to the
database. Reload the file from the database and verify that the
locations are correct. Rollback the database transaction and once
again reload the file. Verify that the original locations are back.
"""
testFileA = File(lfn = "/this/is/a/lfn", size = 1024, events = 10,
checksums = {'cksum':1})
testFileA.addRun(Run( 1, *[45]))
testFileA.create()
testFileA.setLocation(["se1.fnal.gov"])
myThread = threading.currentThread()
myThread.transaction.begin()
testFileA.setLocation(["se1.cern.ch"])
testFileA.setLocation(["bunkse1.fnal.gov", "bunkse1.cern.ch"],
immediateSave = False)
testFileB = File(id = testFileA["id"])
testFileB.loadData()
goldenLocations = ["se1.fnal.gov", "se1.cern.ch"]
for location in testFileB["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
myThread.transaction.rollback()
testFileB.loadData()
goldenLocations = ["se1.fnal.gov"]
for location in testFileB["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
return
def testLoadData(self):
"""
_testLoadData_
Test the loading of all data from a file, including run/lumi
associations, location information and parentage information.
"""
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()
testFileA.setLocation(se="se1.fnal.gov", immediateSave=False)
testFileA.setLocation(se="se1.cern.ch", immediateSave=False)
testFileA.addParent("/this/is/a/parent/lfnA")
testFileA.addParent("/this/is/a/parent/lfnB")
testFileA.updateLocations()
testFileB = File(lfn=testFileA["lfn"])
testFileB.loadData(parentage=1)
testFileC = File(id=testFileA["id"])
testFileC.loadData(parentage=1)
assert testFileA == testFileB, \
"ERROR: File load by LFN didn't work"
assert testFileA == testFileC, \
"ERROR: File load by ID didn't work"
testFileA.delete()
testFileParentA.delete()
testFileParentB.delete()
return
def testDataStructsFile(self):
"""
_testDataStructsFile_
Tests our ability to create a WMBS file from a DataStructs File and vice versa
"""
myThread = threading.currentThread()
testLFN = "lfn1"
testSize = 1024
testEvents = 100
testCksum = {"cksum": '1'}
testParents = set(["lfn2"])
testRun = Run(1, *[45])
testSE = "se1.cern.ch"
parentFile = File(lfn="lfn2")
parentFile.create()
testFile = File()
inputFile = WMFile(lfn=testLFN,
size=testSize,
events=testEvents,
checksums=testCksum,
parents=testParents)
inputFile.addRun(testRun)
inputFile.setLocation(se=testSE)
testFile.loadFromDataStructsFile(file=inputFile)
testFile.create()
testFile.save()
loadFile = File(lfn="lfn1")
loadFile.loadData(parentage=1)
self.assertEqual(loadFile['size'], testSize)
self.assertEqual(loadFile['events'], testEvents)
self.assertEqual(loadFile['checksums'], testCksum)
self.assertEqual(loadFile['locations'], set([testSE]))
#self.assertEqual(loadFile['parents'].pop()['lfn'], 'lfn2')
wmFile = loadFile.returnDataStructsFile()
self.assertEqual(wmFile == inputFile, True)
return
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 testDataStructsFile(self):
"""
_testDataStructsFile_
Tests our ability to create a WMBS file from a DataStructs File and vice versa
"""
myThread = threading.currentThread()
testLFN = "lfn1"
testSize = 1024
testEvents = 100
testCksum = {"cksum": '1'}
testParents = set(["lfn2"])
testRun = Run( 1, *[45])
testSE = "se1.cern.ch"
parentFile = File(lfn= "lfn2")
parentFile.create()
testFile = File()
inputFile = WMFile(lfn = testLFN, size = testSize, events = testEvents, checksums = testCksum, parents = testParents)
inputFile.addRun(testRun)
inputFile.setLocation(se = testSE)
testFile.loadFromDataStructsFile(file = inputFile)
testFile.create()
testFile.save()
loadFile = File(lfn = "lfn1")
loadFile.loadData(parentage = 1)
self.assertEqual(loadFile['size'], testSize)
self.assertEqual(loadFile['events'], testEvents)
self.assertEqual(loadFile['checksums'], testCksum)
self.assertEqual(loadFile['locations'], set([testSE]))
#self.assertEqual(loadFile['parents'].pop()['lfn'], 'lfn2')
wmFile = loadFile.returnDataStructsFile()
self.assertEqual(wmFile == inputFile, True)
return
def testLocationsConstructor(self):
"""
_testLocationsConstructor_
Test to make sure that locations passed into the File() constructor
are loaded from and save to the database correctly. Also test to make
sure that the class behaves well when the location is passed in as a
single string instead of a set.
"""
testFileA = File(lfn = "/this/is/a/lfn", size = 1024, events = 10,
checksums = {'cksum':1}, locations = set(["se1.fnal.gov"]))
testFileA.addRun(Run( 1, *[45]))
testFileA.create()
testFileB = File(lfn = "/this/is/a/lfn2", size = 1024, events = 10,
checksums = {'cksum':1}, locations = "se1.fnal.gov")
testFileB.addRun(Run( 1, *[45]))
testFileB.create()
testFileC = File(id = testFileA["id"])
testFileC.loadData()
goldenLocations = ["se1.fnal.gov"]
for location in testFileC["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
testFileC = File(id = testFileB["id"])
testFileC.loadData()
goldenLocations = ["se1.fnal.gov"]
for location in testFileC["locations"]:
assert location in goldenLocations, \
"ERROR: Unknown file location"
goldenLocations.remove(location)
assert len(goldenLocations) == 0, \
"ERROR: Some locations are missing"
return
def testAddChild(self):
"""
_testAddChild_
Add a child to some parent files and make sure that all the parentage
information is loaded/stored correctly from the database.
"""
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()
testFileParentA.addChild("/this/is/a/lfn")
testFileParentB.addChild("/this/is/a/lfn")
testFileB = File(id=testFileA["id"])
testFileB.loadData(parentage=1)
goldenFiles = [testFileParentA, testFileParentB]
for parentFile in testFileB["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
return
def testCreateWithParent(self):
"""
Test passing parnents arguments in file creation.
check if parent file does not exist, it create the file and set the parentage
"""
# create parent file before it got added to child file.
testFileParentA = File(lfn="/this/is/a/parent/lfnA",
size=1024,
events=20,
checksums={'cksum': 1})
testFileParentA.addRun(Run(1, *[45]))
testFileParentA.create()
# don't create create parent file before it got added to child file.
testFileParentB = File(lfn="/this/is/a/parent/lfnB",
size=1024,
events=20,
checksums={'cksum': 1})
testFileParentB.addRun(Run(1, *[45]))
testFileA = File(lfn="/this/is/a/lfn",
size=1024,
events=10,
checksums={'cksum': 1},
parents=[testFileParentA, testFileParentB])
testFileA.addRun(Run(1, *[45]))
testFileA.create()
testFileB = File(id=testFileA["id"])
testFileB.loadData(parentage=1)
goldenFiles = [testFileParentA, testFileParentB]
for parentFile in testFileB["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
def testLoadData(self):
"""
_testLoadData_
Test the loading of all data from a file, including run/lumi
associations, location information and parentage information.
"""
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()
testFileA.setLocation(se = "se1.fnal.gov", immediateSave = False)
testFileA.setLocation(se = "se1.cern.ch", immediateSave = False)
testFileA.addParent("/this/is/a/parent/lfnA")
testFileA.addParent("/this/is/a/parent/lfnB")
testFileA.updateLocations()
testFileB = File(lfn = testFileA["lfn"])
testFileB.loadData(parentage = 1)
testFileC = File(id = testFileA["id"])
testFileC.loadData(parentage = 1)
assert testFileA == testFileB, \
"ERROR: File load by LFN didn't work"
assert testFileA == testFileC, \
"ERROR: File load by ID didn't work"
testFileA.delete()
testFileParentA.delete()
testFileParentB.delete()
return
def test_SetLocationsForWorkQueue(self):
"""
_SetLocationsForWorkQueue_
Test the code that sets locations for the WorkQueue
This is more complicated then it seems.
"""
action = self.daofactory(classname = "Files.SetLocationForWorkQueue")
testFile = File(lfn = "myLFN", size = 1024,
events = 10, checksums={'cksum':1111})
testFile.create()
tFile1 = File(lfn = "myLFN")
tFile1.loadData()
locations = tFile1.getLocations()
self.assertEqual(locations, [])
binds = [{'lfn': 'myLFN', 'location': 'se1.cern.ch'}]
action.execute(lfns = ['myLFN'], locations = binds)
tFile1.loadData()
locations = tFile1.getLocations()
self.assertEqual(locations, ['se1.cern.ch'])
binds = [{'lfn': 'myLFN', 'location': 'se1.fnal.gov'}]
action.execute(lfns = ['myLFN'], locations = binds)
tFile1.loadData()
locations = tFile1.getLocations()
self.assertEqual(locations, ['se1.fnal.gov'])
return
def testSetLocationByLFN(self):
"""
_testSetLocationByLFN_
Create a file and add a couple locations. Load the file from the
database to make sure that the locations were set correctly.
"""
testFileA = File(lfn="/this/is/a/lfnA",
size=1024,
events=10,
checksums={'cksum': 1})
testFileA.addRun(Run(1, *[45]))
testFileA.create()
testFileB = File(lfn="/this/is/a/lfnB",
size=1024,
events=10,
checksums={'cksum': 1})
testFileB.addRun(Run(1, *[45]))
testFileB.create()
parentAction = self.daofactory(classname="Files.SetLocationByLFN")
binds = [{
'lfn': "/this/is/a/lfnA",
'location': 'se1.fnal.gov'
}, {
'lfn': "/this/is/a/lfnB",
'location': 'se1.fnal.gov'
}]
parentAction.execute(lfn=binds)
testFileC = File(id=testFileA["id"])
testFileC.loadData()
testFileD = File(id=testFileB["id"])
testFileD.loadData()
self.assertEqual(testFileC['locations'], set(['se1.fnal.gov']))
self.assertEqual(testFileD['locations'], set(['se1.fnal.gov']))
return
def testCreateWithParent(self):
"""
Test passing parnents arguments in file creation.
check if parent file does not exist, it create the file and set the parentage
"""
# create parent file before it got added to child file.
testFileParentA = File(lfn = "/this/is/a/parent/lfnA", size = 1024,
events = 20, checksums = {'cksum': 1})
testFileParentA.addRun(Run( 1, *[45]))
testFileParentA.create()
# don't create create parent file before it got added to child file.
testFileParentB = File(lfn = "/this/is/a/parent/lfnB", size = 1024,
events = 20, checksums = {'cksum': 1})
testFileParentB.addRun(Run( 1, *[45]))
testFileA = File(lfn = "/this/is/a/lfn", size = 1024, events = 10,
checksums = {'cksum':1},
parents = [testFileParentA, testFileParentB])
testFileA.addRun(Run( 1, *[45]))
testFileA.create()
testFileB = File(id = testFileA["id"])
testFileB.loadData(parentage = 1)
goldenFiles = [testFileParentA, testFileParentB]
for parentFile in testFileB["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
def test_SetLocationsForWorkQueue(self):
"""
_SetLocationsForWorkQueue_
Test the code that sets locations for the WorkQueue
This is more complicated then it seems.
"""
action = self.daofactory(classname="Files.SetLocationForWorkQueue")
testFile = File(lfn="myLFN",
size=1024,
events=10,
checksums={'cksum': 1111})
testFile.create()
tFile1 = File(lfn="myLFN")
tFile1.loadData()
locations = tFile1.getLocations()
self.assertEqual(locations, [])
binds = [{'lfn': 'myLFN', 'location': 'se1.cern.ch'}]
action.execute(lfns=['myLFN'], locations=binds)
tFile1.loadData()
locations = tFile1.getLocations()
self.assertEqual(locations, ['se1.cern.ch'])
binds = [{'lfn': 'myLFN', 'location': 'se1.fnal.gov'}]
action.execute(lfns=['myLFN'], locations=binds)
tFile1.loadData()
locations = tFile1.getLocations()
self.assertEqual(locations, ['se1.fnal.gov'])
return
if newfile.exists() == False:
newfile.create()
filesetToProcess.addFile(newfile)
filesetToProcess.setLastUpdate(\
int(time.time()))
filesetToProcess.commit()
runSet = set()
runSet.add(Run( lumi\
['RunNumber' ], *[lumi['LumiSectionNumber']] ))
newfile.addRunSet(runSet)
else:
newfile.loadData()
listFile = fileInFileset.execute\
(filesetToProcess.id)
if {'fileid': newfile[\
'id']} not in listFile:
filesetToProcess.addFile(newfile)
filesetToProcess.setLastUpdate\
(int(time.time()))
filesetToProcess.commit()
val = 0
for run in newfile['runs']:
if lumi['RunNumber'] == run.run:
def testAddChildTransaction(self):
"""
_testAddChildTransaction_
Add a child to some parent files and make sure that all the parentage
information is loaded/stored correctly from the database. Rollback the
addition of one of the childs and then verify that it does in fact only
have one parent.
"""
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()
testFileParentA.addChild("/this/is/a/lfn")
myThread = threading.currentThread()
myThread.transaction.begin()
testFileParentB.addChild("/this/is/a/lfn")
testFileB = File(id=testFileA["id"])
testFileB.loadData(parentage=1)
goldenFiles = [testFileParentA, testFileParentB]
for parentFile in testFileB["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
myThread.transaction.rollback()
testFileB.loadData(parentage=1)
goldenFiles = [testFileParentA]
for parentFile in testFileB["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
return
def testBulkParentage(self):
"""
_testBulkParentage_
Verify that the bulk parentage dao correctly sets file parentage.
"""
testFileChildA = File(lfn="/this/is/a/child/lfnA",
size=1024,
events=20,
checksums={'cksum': 1})
testFileChildB = File(lfn="/this/is/a/child/lfnB",
size=1024,
events=20,
checksums={'cksum': 1})
testFileChildA.create()
testFileChildB.create()
testFileA = File(lfn="/this/is/a/lfnA",
size=1024,
events=10,
checksums={'cksum': 1})
testFileA.create()
testFileB = File(lfn="/this/is/a/lfnB",
size=1024,
events=10,
checksums={'cksum': 1})
testFileB.create()
testFileC = File(lfn="/this/is/a/lfnC",
size=1024,
events=10,
checksums={'cksum': 1})
testFileC.create()
parentage = [{
"child": testFileChildA["id"],
"parent": testFileA["id"]
}, {
"child": testFileChildA["id"],
"parent": testFileB["id"]
}, {
"child": testFileChildA["id"],
"parent": testFileC["id"]
}, {
"child": testFileChildB["id"],
"parent": testFileA["id"]
}, {
"child": testFileChildB["id"],
"parent": testFileB["id"]
}]
myThread = threading.currentThread()
daofactory = DAOFactory(package="WMCore.WMBS",
logger=myThread.logger,
dbinterface=myThread.dbi)
bulkParentageAction = daofactory(classname="Files.AddBulkParentage")
bulkParentageAction.execute(parentage)
testFileD = File(id=testFileChildA["id"])
testFileD.loadData(parentage=1)
testFileE = File(id=testFileChildB["id"])
testFileE.loadData(parentage=1)
goldenFiles = [testFileA, testFileB, testFileC]
for parentFile in testFileD["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
goldenFiles = [testFileA, testFileB]
for parentFile in testFileE["parents"]:
assert parentFile in goldenFiles, \
"ERROR: Unknown parent file"
goldenFiles.remove(parentFile)
assert len(goldenFiles) == 0, \
"ERROR: Some parents are missing"
return
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 T0ASTFile(dict):
"""
_T0ASTFile_
A dictionary based object meant to represent a file in T0AST.
It contains the following keys:
ID
STREAM
RUN_ID
FILESIZE
EVENTS
CKSUM
LFN
DATASET_ID
DATA_TIER
TYPE
STATUS
EXPORT_STATUS
STREAM
STREAM_ID
LUMI_ID
LUMI_LIST
PARENT_LIST
PSET_HASH
BRANCH_HASH
LOCATIONS
"""
def __init__(self, wmbsFile=None, **args):
"""
___init___
Initialize all attributes.
"""
dict.__init__(self)
self.runLumi = None
# place holder for list of parent(repacked) files names
# only used for reco type, for repacked this should be empty
self.parentList = None
self.parentIDList = None
# this is the dataset_path_ID in data table not primary dataset id
self.datasetPathID = None
# place holder for PSet hash, Gets from Job Report
self.setdefault("RUN_ID", None)
self.setdefault("STREAM", None)
self.setdefault("DATASET_ID", None)
self.setdefault("PRIMARY_DATASET", None)
self.setdefault("PROCESSED_DATASET", None)
self.setdefault("DATA_TIER", None)
self.setdefault("BLOCK_ID", None)
self.setdefault("STREAM", None)
self.setdefault("STREAM_ID", None)
# currently using WM Run structure here but need to use lumi structure
# place holder for list of lumi data struct
#self.setdefault("LUMI_LIST", [])
self.setdefault("PSET_HASH", None)
self.setdefault("BRANCH_HASH", None)
self.setdefault("LOCATIONS", None)
self.wmbsFile = wmbsFile
if wmbsFile == None:
self.setdefault("ID", None)
self.setdefault("FILESIZE", None)
self.setdefault("EVENTS", None)
self.setdefault("FIRST_EVENT", 0)
self.setdefault("LAST_EVENT", 0)
self.setdefault("CKSUM", None)
self.setdefault("LFN", None)
# this is primary dataset ID
self.update(args)
else:
self["ID"] = wmbsFile["id"]
self["FILESIZE"] = wmbsFile["size"]
self["CKSUM"] = wmbsFile["cksum"]
self["EVENTS"] = wmbsFile["events"]
self["FIRST_EVENT"] = wmbsFile["first_event"]
self["LAST_EVENT"] = wmbsFile["last_event"]
self["LFN"] = wmbsFile["lfn"]
def convertToWMBSFile(self):
"""
__getWMBSFile__
return WMBSFile instance converted from T0AST file.
"""
if self.wmbsFile == None:
# hack: make sure first_event/last_event is None
# if they are add 0 event
if self["FIRST_EVENT"] == None:
self["FIRST_EVENT"] = 0
if self["LAST_EVENT"] == None:
self["LAST_EVENT"] = 0
self.wmbsFile = WMBSFile(lfn=self["LFN"], size=self["FILESIZE"],
events=self["EVENTS"], cksum=self["CKSUM"],
first_event = self["FIRST_EVENT"],
last_event = self["LAST_EVENT"],
locations=self["LOCATIONS"])
self.wmbsFile.create()
self["ID"] = self.wmbsFile["id"]
return self.wmbsFile
def getParentList(self, dataType='lfn', stream="Normal"):
"""
_getParentList_
"""
if self["DATA_TIER"] == "RAW":
return []
#TODO: place holder for correct express stream handling
if stream == "Express":
# TODO get using the lumi comparason
# For now return empty list for parent
return []
self.convertToWMBSFile()
if dataType == "id":
parents = self.parentIDList
elif dataType == "lfn":
parents = self.parentList
else:
raise Exception, "Unknown Type"
if parents == None:
# this will retrun right parents for both RECO -> RAW and
# RAW -> [] files except the big files merged directly
if self["DATA_TIER"] == "ALCARECO" \
and alcaNewSelection():
parents = self.wmbsFile.getAncestors(
level=3, type=dataType)
else:
parents = self.wmbsFile.getAncestors(level=2, type=dataType)
#update the private variable to prevent expensive call over and over again
if dataType == "id":
self.parentIDList = parents
elif dataType == "lfn":
self.parentList = parents
return parents
def getLumiList(self):
"""
_getLumiList_
"""
self.convertToWMBSFile()
if self.runLumi == None:
if len(self.wmbsFile["runs"]) == 0:
self.wmbsFile.loadData()
self.runLumi = list(self.wmbsFile["runs"])[0]
self["RUN_ID"] = self.runLumi.run
return self.runLumi
def getRunID(self):
"""
_getRunID_
"""
self.convertToWMBSFile()
if self["RUN_ID"] == None:
if len(self.wmbsFile["runs"]) == 0:
self.wmbsFile.loadData()
self.runLumi = list(self.wmbsFile["runs"])[0]
self["RUN_ID"] = self.runLumi.run
return self["RUN_ID"]
def getDatasetPathID(self):
"""
_getDatasetPathID_
To do: this need database connection to get the value.
need to be modified like wmbs wrapper class
"""
return self.datasetPathID
def __call__(self, filesetToProcess):
"""
The algorithm itself
"""
global LOCK
# Get configuration
initObj = WMInit()
initObj.setLogging()
initObj.setDatabaseConnection(os.getenv("DATABASE"), \
os.getenv('DIALECT'), os.getenv("DBSOCK"))
myThread = threading.currentThread()
daofactory = DAOFactory(package = "WMCore.WMBS" , \
logger = myThread.logger, \
dbinterface = myThread.dbi)
locationNew = daofactory(classname = "Locations.New")
getFileLoc = daofactory(classname = "Files.GetLocation")
logging.debug("the T0Feeder is processing %s" % \
filesetToProcess.name)
logging.debug("the fileset name %s" % \
(filesetToProcess.name).split(":")[0])
startRun = (filesetToProcess.name).split(":")[3]
fileType = (filesetToProcess.name).split(":")[2]
# url builder
primaryDataset = ((filesetToProcess.name).split(":")[0]).split('/')[1]
processedDataset = ((filesetToProcess.name).split(":")[0]).split('/')[2]
dataTier = (((filesetToProcess.name\
).split(":")[0]).split('/')[3]).split('-')[0]
# Fisrt call to T0 db for this fileset
# Here add test for the closed fileset
LASTIME = filesetToProcess.lastUpdate
url = "/tier0/listfilesoverinterval/%s/%s/%s/%s/%s" % \
(fileType, LASTIME, primaryDataset,processedDataset, dataTier)
tries = 1
while True:
try:
myRequester = JSONRequests(url = "vocms52.cern.ch:8889")
requestResult = myRequester.get(\
url+"/"+"?return_type=text/json%2Bdas")
newFilesList = requestResult[0]["results"]
except:
logging.debug("T0Reader call error...")
if tries == self.maxRetries:
return
else:
tries += 1
continue
logging.debug("T0ASTRun queries done ...")
now = time.time()
filesetToProcess.last_update = now
LASTIME = int(newFilesList['end_time']) + 1
break
# process all files
if len(newFilesList['files']):
LOCK.acquire()
try:
locationNew.execute(siteName = "caf.cern.ch", seName = "caf.cern.ch")
except Exception as e:
logging.debug("Error when adding new location...")
logging.debug(e)
logging.debug( format_exc() )
for files in newFilesList['files']:
# Assume parents aren't asked
newfile = File(str(files['lfn']), \
size = files['file_size'], events = files['events'])
try:
if newfile.exists() == False :
newfile.create()
else:
newfile.loadData()
#Add run test if already exist
for run in files['runs']:
if startRun != 'None' and int(startRun) <= int(run):
# ToDo: Distinguish between
# filestA-RunX and filesetA-Run[0-9]*
filesetRun = Fileset( name = (((\
filesetToProcess.name).split(':')[0]).split('/')[0]\
)+'/'+(((filesetToProcess.name).split(':')[0]).split\
('/')[1])+'/'+(((filesetToProcess.name).split(':')[0]\
).split('/')[2])+'/'+((((filesetToProcess.name).split\
(':')[0]).split('/')[3]).split('-')[0])+'-'+'Run'+str\
(run)+":"+":".join((filesetToProcess.name).split(':')[1:] \
) )
if filesetRun.exists() == False :
filesetRun.create()
else:
filesetRun.loadData()
# Add test runs already there
# (for growing dataset) -
# to support file with different runs and lumi
if not newfile['runs']:
runSet = set()
runSet.add(Run( run, *files['runs'][run]))
newfile.addRunSet(runSet)
fileLoc = getFileLoc.execute(file = files['lfn'])
if 'caf.cern.ch' not in fileLoc:
newfile.setLocation("caf.cern.ch")
filesetRun.addFile(newfile)
logging.debug("new file created/loaded added by T0ASTRun...")
filesetRun.commit()
except Exception as e:
logging.debug("Error when adding new files in T0ASTRun...")
logging.debug(e)
logging.debug( format_exc() )
filesetToProcess.setLastUpdate\
(int(newFilesList['end_time']) + 1)
filesetToProcess.commit()
LOCK.release()
else:
logging.debug("nothing to do...")
# For re-opned fileset or empty, try until the purge time
if (int(now)/3600 - LASTIME/3600) > self.reopenTime:
filesetToProcess.setLastUpdate(time.time())
filesetToProcess.commit()
if LASTIME:
myRequester = JSONRequests(url = "vocms52.cern.ch:8889")
requestResult = myRequester.get("/tier0/runs")
for listRun in requestResult[0]:
if int(startRun) <= int(listRun['run']):
if listRun['status'] =='CloseOutExport' or \
listRun['status'] =='Complete' or listRun['status'] ==\
'CloseOutT1Skimming':
closeFileset = Fileset( name = (((\
filesetToProcess.name).split(':')[0]).split('/')[0])+'/'+\
(((filesetToProcess.name).split(':')[0]).split('/')[1]\
)+'/'+(((filesetToProcess.name).split(':')[0]).split('/')\
[2])+'/'+((((filesetToProcess.name).split(':')[0]).split\
('/')[3]).split('-')[0])+'-'+'Run'+str(listRun['run'])\
+":"+":".join((filesetToProcess.name).split(':')[1:] ) )
if closeFileset.exists() != False :
closeFileset = Fileset( id = closeFileset.exists())
closeFileset.loadData()
if closeFileset.open == True:
closeFileset.markOpen(False)
# Commit the fileset
filesetToProcess.commit()
# Commit the fileset
logging.debug("Test purge in T0ASTRun ...")
filesetToProcess.load()
LASTIME = filesetToProcess.lastUpdate
if (int(now)/3600 - LASTIME/3600) > self.purgeTime:
filesetToProcess.markOpen(False)
logging.debug("Purge Done...")
filesetToProcess.commit()
def __call__(self, filesetToProcess):
"""
The algorithm itself
"""
global LOCK
# Get configuration
initObj = WMInit()
initObj.setLogging()
initObj.setDatabaseConnection(os.getenv("DATABASE"), \
os.getenv('DIALECT'), os.getenv("DBSOCK"))
myThread = threading.currentThread()
daofactory = DAOFactory(package = "WMCore.WMBS" , \
logger = myThread.logger, \
dbinterface = myThread.dbi)
locationNew = daofactory(classname="Locations.New")
getFileLoc = daofactory(classname="Files.GetLocation")
fileInFileset = daofactory(classname="Files.InFileset")
logging.debug("the T0Feeder is processing %s" % \
filesetToProcess.name)
logging.debug("the fileset name %s" % \
(filesetToProcess.name).split(":")[0])
# Get the start Run if asked
startRun = (filesetToProcess.name).split(":")[3]
fileType = (filesetToProcess.name).split(":")[2]
LASTIME = filesetToProcess.lastUpdate
# url builder
primaryDataset = ((filesetToProcess.name).split(":")[0]).split('/')[1]
processedDataset = ((
filesetToProcess.name).split(":")[0]).split('/')[2]
dataTier = ((filesetToProcess.name\
).split(":")[0]).split('/')[3]
url = "/tier0/listfilesoverinterval/%s/%s/%s/%s/%s" % \
(fileType, LASTIME, primaryDataset,processedDataset, dataTier)
tries = 1
while True:
try:
myRequester = JSONRequests(url="vocms52.cern.ch:8889")
requestResult = myRequester.get(\
url+"/"+"?return_type=text/json%2Bdas")
newFilesList = requestResult[0]["results"]
except:
logging.debug("T0Reader call error...")
if tries == self.maxRetries:
return
else:
tries += 1
continue
logging.debug("T0 queries done ...")
now = time.time()
LASTIME = int(newFilesList['end_time']) + 1
break
# process all files
if len(newFilesList['files']):
try:
locationNew.execute(siteName="caf.cern.ch",
seName="caf.cern.ch")
except Exception as e:
logging.debug("Error when adding new location...")
logging.debug(e)
logging.debug(format_exc())
for files in newFilesList['files']:
# Assume parents aren't asked
newfile = File(str(files['lfn']), \
size = files['file_size'], events = files['events'])
try:
LOCK.acquire()
if newfile.exists() == False:
newfile.create()
for run in files['runs']:
runSet = set()
runSet.add(Run(run, *files['runs'][run]))
newfile.addRunSet(runSet)
else:
newfile.loadData()
fileLoc = getFileLoc.execute(file=files['lfn'])
if 'caf.cern.ch' not in fileLoc:
newfile.setLocation("caf.cern.ch")
# else:
# logging.debug("File already associated to %s" %fileLoc)
LOCK.release()
if len(newfile["runs"]):
val = 0
for run in newfile['runs']:
if run.run < int(startRun):
val = 1
break
if not val:
listFile = fileInFileset.execute\
(filesetToProcess.id)
if {'fileid': newfile['id']} not in listFile:
filesetToProcess.addFile(newfile)
filesetToProcess.setLastUpdate\
(int(newFilesList['end_time']) + 1)
filesetToProcess.commit()
logging.debug(
"new file created/loaded added by T0AST..."
)
except Exception as e:
logging.debug("Error when adding new files in T0AST...")
logging.debug(e)
logging.debug(format_exc())
LOCK.release()
filesetToProcess.commit()
else:
logging.debug("nothing to do in T0AST...")
# For reopned fileset or empty
# try until the purge time is reached
if (int(now) / 3600 - LASTIME / 3600) > self.reopenTime:
filesetToProcess.setLastUpdate(time.time())
filesetToProcess.commit()
# Commit the fileset
logging.debug("Test purge in T0AST ...")
filesetToProcess.load()
LASTIME = filesetToProcess.lastUpdate
if (int(now) / 3600 - LASTIME / 3600) > self.purgeTime:
filesetToProcess.markOpen(False)
logging.debug("Purge Done...")
filesetToProcess.commit()
def __call__(self, filesetToProcess):
"""
The algorithm itself
"""
global LOCK
# Get configuration
initObj = WMInit()
initObj.setLogging()
initObj.setDatabaseConnection(os.getenv("DATABASE"), \
os.getenv('DIALECT'), os.getenv("DBSOCK"))
myThread = threading.currentThread()
daofactory = DAOFactory(package = "WMCore.WMBS" , \
logger = myThread.logger, \
dbinterface = myThread.dbi)
locationNew = daofactory(classname="Locations.New")
getFileLoc = daofactory(classname="Files.GetLocation")
logging.debug("the T0Feeder is processing %s" % \
filesetToProcess.name)
logging.debug("the fileset name %s" % \
(filesetToProcess.name).split(":")[0])
startRun = (filesetToProcess.name).split(":")[3]
fileType = (filesetToProcess.name).split(":")[2]
# url builder
primaryDataset = ((filesetToProcess.name).split(":")[0]).split('/')[1]
processedDataset = ((
filesetToProcess.name).split(":")[0]).split('/')[2]
dataTier = (((filesetToProcess.name\
).split(":")[0]).split('/')[3]).split('-')[0]
# Fisrt call to T0 db for this fileset
# Here add test for the closed fileset
LASTIME = filesetToProcess.lastUpdate
url = "/tier0/listfilesoverinterval/%s/%s/%s/%s/%s" % \
(fileType, LASTIME, primaryDataset,processedDataset, dataTier)
tries = 1
while True:
try:
myRequester = JSONRequests(url="vocms52.cern.ch:8889")
requestResult = myRequester.get(\
url+"/"+"?return_type=text/json%2Bdas")
newFilesList = requestResult[0]["results"]
except:
logging.debug("T0Reader call error...")
if tries == self.maxRetries:
return
else:
tries += 1
continue
logging.debug("T0ASTRun queries done ...")
now = time.time()
filesetToProcess.last_update = now
LASTIME = int(newFilesList['end_time']) + 1
break
# process all files
if len(newFilesList['files']):
LOCK.acquire()
try:
locationNew.execute(siteName="caf.cern.ch",
seName="caf.cern.ch")
except Exception as e:
logging.debug("Error when adding new location...")
logging.debug(e)
logging.debug(format_exc())
for files in newFilesList['files']:
# Assume parents aren't asked
newfile = File(str(files['lfn']), \
size = files['file_size'], events = files['events'])
try:
if newfile.exists() == False:
newfile.create()
else:
newfile.loadData()
#Add run test if already exist
for run in files['runs']:
if startRun != 'None' and int(startRun) <= int(run):
# ToDo: Distinguish between
# filestA-RunX and filesetA-Run[0-9]*
filesetRun = Fileset( name = (((\
filesetToProcess.name).split(':')[0]).split('/')[0]\
)+'/'+(((filesetToProcess.name).split(':')[0]).split\
('/')[1])+'/'+(((filesetToProcess.name).split(':')[0]\
).split('/')[2])+'/'+((((filesetToProcess.name).split\
(':')[0]).split('/')[3]).split('-')[0])+'-'+'Run'+str\
(run)+":"+":".join((filesetToProcess.name).split(':')[1:] \
) )
if filesetRun.exists() == False:
filesetRun.create()
else:
filesetRun.loadData()
# Add test runs already there
# (for growing dataset) -
# to support file with different runs and lumi
if not newfile['runs']:
runSet = set()
runSet.add(Run(run, *files['runs'][run]))
newfile.addRunSet(runSet)
fileLoc = getFileLoc.execute(file=files['lfn'])
if 'caf.cern.ch' not in fileLoc:
newfile.setLocation("caf.cern.ch")
filesetRun.addFile(newfile)
logging.debug(
"new file created/loaded added by T0ASTRun...")
filesetRun.commit()
except Exception as e:
logging.debug("Error when adding new files in T0ASTRun...")
logging.debug(e)
logging.debug(format_exc())
filesetToProcess.setLastUpdate\
(int(newFilesList['end_time']) + 1)
filesetToProcess.commit()
LOCK.release()
else:
logging.debug("nothing to do...")
# For re-opned fileset or empty, try until the purge time
if (int(now) / 3600 - LASTIME / 3600) > self.reopenTime:
filesetToProcess.setLastUpdate(time.time())
filesetToProcess.commit()
if LASTIME:
myRequester = JSONRequests(url="vocms52.cern.ch:8889")
requestResult = myRequester.get("/tier0/runs")
for listRun in requestResult[0]:
if int(startRun) <= int(listRun['run']):
if listRun['status'] =='CloseOutExport' or \
listRun['status'] =='Complete' or listRun['status'] ==\
'CloseOutT1Skimming':
closeFileset = Fileset( name = (((\
filesetToProcess.name).split(':')[0]).split('/')[0])+'/'+\
(((filesetToProcess.name).split(':')[0]).split('/')[1]\
)+'/'+(((filesetToProcess.name).split(':')[0]).split('/')\
[2])+'/'+((((filesetToProcess.name).split(':')[0]).split\
('/')[3]).split('-')[0])+'-'+'Run'+str(listRun['run'])\
+":"+":".join((filesetToProcess.name).split(':')[1:] ) )
if closeFileset.exists() != False:
closeFileset = Fileset(id=closeFileset.exists())
closeFileset.loadData()
if closeFileset.open == True:
closeFileset.markOpen(False)
# Commit the fileset
filesetToProcess.commit()
# Commit the fileset
logging.debug("Test purge in T0ASTRun ...")
filesetToProcess.load()
LASTIME = filesetToProcess.lastUpdate
if (int(now) / 3600 - LASTIME / 3600) > self.purgeTime:
filesetToProcess.markOpen(False)
logging.debug("Purge Done...")
filesetToProcess.commit()
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
if newfile.exists() == False :
newfile.create()
filesetToProcess.addFile(newfile)
filesetToProcess.setLastUpdate(\
int(time.time()))
filesetToProcess.commit()
runSet = set()
runSet.add(Run( lumi\
['RunNumber' ], *[lumi['LumiSectionNumber']] ))
newfile.addRunSet(runSet)
else:
newfile.loadData()
listFile = fileInFileset.execute\
(filesetToProcess.id)
if {'fileid': newfile[\
'id']} not in listFile:
filesetToProcess.addFile(newfile)
filesetToProcess.setLastUpdate\
(int(time.time()))
filesetToProcess.commit()
val = 0
for run in newfile['runs']:
if lumi['RunNumber' ] == run.run:
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"].keys():
assert ckType in outputFile["checksums"].keys(), \
"Error: Output file is missing checksums: %s" % ckType
assert outputFile["checksums"][ckType] == fwkJobReportFile["checksums"][ckType], \
"Error: Checksums don't match."
assert len(fwkJobReportFile["checksums"].keys()) == \
len(outputFile["checksums"].keys()), \
"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.keys()) == 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 __call__(self, filesetToProcess):
"""
The algorithm itself
"""
# Get configuration
initObj = WMInit()
initObj.setLogging()
initObj.setDatabaseConnection(os.getenv("DATABASE"), \
os.getenv('DIALECT'), os.getenv("DBSOCK"))
myThread = threading.currentThread()
daofactory = DAOFactory(package = "WMCore.WMBS" , \
logger = myThread.logger, \
dbinterface = myThread.dbi)
lastFileset = daofactory(classname = "Fileset.ListFilesetByTask")
lastWorkflow = daofactory(classname = "Workflow.LoadFromTask")
subsRun = daofactory(\
classname = "Subscriptions.LoadFromFilesetWorkflow")
successJob = daofactory(classname = "Subscriptions.SucceededJobs")
allJob = daofactory(classname = "Subscriptions.Jobs")
fileInFileset = daofactory(classname = "Files.InFileset")
# Get the start Run if asked
startRun = (filesetToProcess.name).split(":")[3]
logging.debug("the T0Feeder is processing %s" % \
filesetToProcess.name)
logging.debug("the fileset name %s" % \
(filesetToProcess.name).split(":")[0])
fileType = (filesetToProcess.name).split(":")[2]
crabTask = filesetToProcess.name.split(":")[0]
LASTIME = filesetToProcess.lastUpdate
tries = 1
while True:
try:
myRequester = JSONRequests(url = "vocms52.cern.ch:8889")
requestResult = myRequester.get("/tier0/runs")
except:
logging.debug("T0Reader call error...")
if tries == self.maxRetries:
return
else:
tries += 1
continue
logging.debug("T0ASTRunChain feeder queries done ...")
now = time.time()
break
for listRun in requestResult[0]:
if startRun != 'None' and int(listRun['run']) >= int(startRun):
if listRun['status'] =='CloseOutExport' or listRun\
['status']=='Complete' or listRun['status']=='CloseOutT1Skimming':
crabWorkflow = lastWorkflow.execute(task=crabTask)
crabFileset = lastFileset.execute\
(task=crabTask)
crabrunFileset = Fileset(\
name = crabFileset[0]["name"].split(':')[0].split\
('-Run')[0]+ '-Run' + str(listRun['run']) + ":" + \
":".join(crabFileset[0]['name'].split(':')[1:]) )
if crabrunFileset.exists() > 0:
crabrunFileset.load()
currSubs = subsRun.execute\
(crabrunFileset.id, crabWorkflow[0]['id'])
if currSubs:
listsuccessJob = successJob.execute(\
subscription=currSubs['id'])
listallJob = allJob.execute(\
subscription=currSubs['id'])
if len(listsuccessJob) == len(listallJob):
for currid in listsuccessJob:
currjob = Job( id = currid )
currjob.load()
logging.debug("Reading FJR %s" %currjob['fwjr_path'])
jobReport = readJobReport(currjob['fwjr_path'])
if len(jobReport) > 0:
if jobReport[0].files:
for newFile in jobReport[0].files:
logging.debug(\
"Output path %s" %newFile['LFN'])
newFileToAdd = File(\
lfn=newFile['LFN'], locations ='caf.cern.ch')
LOCK.acquire()
if newFileToAdd.exists\
() == False :
newFileToAdd.create()
else:
newFileToAdd.loadData()
LOCK.release()
listFile = \
fileInFileset.execute(filesetToProcess.id)
if {'fileid': \
newFileToAdd['id']} not in listFile:
filesetToProcess.addFile(\
newFileToAdd)
filesetToProcess\
.setLastUpdate(now)
filesetToProcess.commit()
logging.debug(\
"new file created/loaded and added by T0ASTRunChain...")
elif jobReport[0].analysisFiles:
for newFile in jobReport\
[0].analysisFiles:
logging.debug(\
"Ouput path %s " %newFile['LFN'])
newFileToAdd = File(\
lfn=newFile['LFN'], locations ='caf.cern.ch')
LOCK.acquire()
if newFileToAdd.exists\
() == False :
newFileToAdd.create()
else:
newFileToAdd.loadData()
LOCK.release()
listFile = \
fileInFileset.execute(filesetToProcess.id)
if {'fileid': newFileToAdd\
['id']} not in listFile:
logging.debug\
("%s loaded and added by T0ASTRunChain" %newFile['LFN'])
filesetToProcess.addFile\
(newFileToAdd)
filesetToProcess.\
setLastUpdate(now)
filesetToProcess.commit()
logging.debug(\
"new file created/loaded added by T0ASTRunChain...")
else: break #Missed fjr - Try next time
# Commit the fileset
logging.debug("Test purge in T0ASTRunChain ...")
filesetToProcess.load()
LASTIME = filesetToProcess.lastUpdate
# For re-opned fileset or empty, try until the purge time
if (int(now)/3600 - LASTIME/3600) > self.reopenTime:
filesetToProcess.setLastUpdate(time.time())
filesetToProcess.commit()
if (int(now)/3600 - LASTIME/3600) > self.purgeTime:
filesetToProcess.markOpen(False)
logging.debug("Purge Done...")
def __call__(self, filesetToProcess):
"""
The algorithm itself
"""
# Get configuration
initObj = WMInit()
initObj.setLogging()
initObj.setDatabaseConnection(os.getenv("DATABASE"), \
os.getenv('DIALECT'), os.getenv("DBSOCK"))
myThread = threading.currentThread()
daofactory = DAOFactory(package = "WMCore.WMBS" , \
logger = myThread.logger, \
dbinterface = myThread.dbi)
lastFileset = daofactory(classname="Fileset.ListFilesetByTask")
lastWorkflow = daofactory(classname="Workflow.LoadFromTask")
subsRun = daofactory(\
classname = "Subscriptions.LoadFromFilesetWorkflow")
successJob = daofactory(classname="Subscriptions.SucceededJobs")
allJob = daofactory(classname="Subscriptions.Jobs")
fileInFileset = daofactory(classname="Files.InFileset")
# Get the start Run if asked
startRun = (filesetToProcess.name).split(":")[3]
logging.debug("the T0Feeder is processing %s" % \
filesetToProcess.name)
logging.debug("the fileset name %s" % \
(filesetToProcess.name).split(":")[0])
fileType = (filesetToProcess.name).split(":")[2]
crabTask = filesetToProcess.name.split(":")[0]
LASTIME = filesetToProcess.lastUpdate
tries = 1
while True:
try:
myRequester = JSONRequests(url="vocms52.cern.ch:8889")
requestResult = myRequester.get("/tier0/runs")
except:
logging.debug("T0Reader call error...")
if tries == self.maxRetries:
return
else:
tries += 1
continue
logging.debug("T0ASTRunChain feeder queries done ...")
now = time.time()
break
for listRun in requestResult[0]:
if startRun != 'None' and int(listRun['run']) >= int(startRun):
if listRun['status'] =='CloseOutExport' or listRun\
['status']=='Complete' or listRun['status']=='CloseOutT1Skimming':
crabWorkflow = lastWorkflow.execute(task=crabTask)
crabFileset = lastFileset.execute\
(task=crabTask)
crabrunFileset = Fileset(\
name = crabFileset[0]["name"].split(':')[0].split\
('-Run')[0]+ '-Run' + str(listRun['run']) + ":" + \
":".join(crabFileset[0]['name'].split(':')[1:]) )
if crabrunFileset.exists() > 0:
crabrunFileset.load()
currSubs = subsRun.execute\
(crabrunFileset.id, crabWorkflow[0]['id'])
if currSubs:
listsuccessJob = successJob.execute(\
subscription=currSubs['id'])
listallJob = allJob.execute(\
subscription=currSubs['id'])
if len(listsuccessJob) == len(listallJob):
for currid in listsuccessJob:
currjob = Job(id=currid)
currjob.load()
logging.debug("Reading FJR %s" %
currjob['fwjr_path'])
jobReport = readJobReport(
currjob['fwjr_path'])
if len(jobReport) > 0:
if jobReport[0].files:
for newFile in jobReport[0].files:
logging.debug(\
"Output path %s" %newFile['LFN'])
newFileToAdd = File(\
lfn=newFile['LFN'], locations ='caf.cern.ch')
LOCK.acquire()
if newFileToAdd.exists\
() == False :
newFileToAdd.create()
else:
newFileToAdd.loadData()
LOCK.release()
listFile = \
fileInFileset.execute(filesetToProcess.id)
if {'fileid': \
newFileToAdd['id']} not in listFile:
filesetToProcess.addFile(\
newFileToAdd)
filesetToProcess\
.setLastUpdate(now)
filesetToProcess.commit()
logging.debug(\
"new file created/loaded and added by T0ASTRunChain...")
elif jobReport[0].analysisFiles:
for newFile in jobReport\
[0].analysisFiles:
logging.debug(\
"Ouput path %s " %newFile['LFN'])
newFileToAdd = File(\
lfn=newFile['LFN'], locations ='caf.cern.ch')
LOCK.acquire()
if newFileToAdd.exists\
() == False :
newFileToAdd.create()
else:
newFileToAdd.loadData()
LOCK.release()
listFile = \
fileInFileset.execute(filesetToProcess.id)
if {'fileid': newFileToAdd\
['id']} not in listFile:
logging.debug\
("%s loaded and added by T0ASTRunChain" %newFile['LFN'])
filesetToProcess.addFile\
(newFileToAdd)
filesetToProcess.\
setLastUpdate(now)
filesetToProcess.commit()
logging.debug(\
"new file created/loaded added by T0ASTRunChain...")
else:
break #Missed fjr - Try next time
# Commit the fileset
logging.debug("Test purge in T0ASTRunChain ...")
filesetToProcess.load()
LASTIME = filesetToProcess.lastUpdate
# For re-opned fileset or empty, try until the purge time
if (int(now) / 3600 - LASTIME / 3600) > self.reopenTime:
filesetToProcess.setLastUpdate(time.time())
filesetToProcess.commit()
if (int(now) / 3600 - LASTIME / 3600) > self.purgeTime:
filesetToProcess.markOpen(False)
logging.debug("Purge Done...")
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 __call__(self, filesetToProcess):
"""
The algorithm itself
"""
global LOCK
# Get configuration
initObj = WMInit()
initObj.setLogging()
initObj.setDatabaseConnection(os.getenv("DATABASE"), os.getenv("DIALECT"), os.getenv("DBSOCK"))
myThread = threading.currentThread()
daofactory = DAOFactory(package="WMCore.WMBS", logger=myThread.logger, dbinterface=myThread.dbi)
locationNew = daofactory(classname="Locations.New")
getFileLoc = daofactory(classname="Files.GetLocation")
fileInFileset = daofactory(classname="Files.InFileset")
logging.debug("the T0Feeder is processing %s" % filesetToProcess.name)
logging.debug("the fileset name %s" % (filesetToProcess.name).split(":")[0])
# Get the start Run if asked
startRun = (filesetToProcess.name).split(":")[3]
fileType = (filesetToProcess.name).split(":")[2]
LASTIME = filesetToProcess.lastUpdate
# url builder
primaryDataset = ((filesetToProcess.name).split(":")[0]).split("/")[1]
processedDataset = ((filesetToProcess.name).split(":")[0]).split("/")[2]
dataTier = ((filesetToProcess.name).split(":")[0]).split("/")[3]
url = "/tier0/listfilesoverinterval/%s/%s/%s/%s/%s" % (
fileType,
LASTIME,
primaryDataset,
processedDataset,
dataTier,
)
tries = 1
while True:
try:
myRequester = JSONRequests(url="vocms52.cern.ch:8889")
requestResult = myRequester.get(url + "/" + "?return_type=text/json%2Bdas")
newFilesList = requestResult[0]["results"]
except:
logging.debug("T0Reader call error...")
if tries == self.maxRetries:
return
else:
tries += 1
continue
logging.debug("T0 queries done ...")
now = time.time()
LASTIME = int(newFilesList["end_time"]) + 1
break
# process all files
if len(newFilesList["files"]):
try:
locationNew.execute(siteName="caf.cern.ch", seName="caf.cern.ch")
except Exception, e:
logging.debug("Error when adding new location...")
logging.debug(e)
logging.debug(format_exc())
for files in newFilesList["files"]:
# Assume parents aren't asked
newfile = File(str(files["lfn"]), size=files["file_size"], events=files["events"])
try:
LOCK.acquire()
if newfile.exists() == False:
newfile.create()
for run in files["runs"]:
runSet = set()
runSet.add(Run(run, *files["runs"][run]))
newfile.addRunSet(runSet)
else:
newfile.loadData()
fileLoc = getFileLoc.execute(file=files["lfn"])
if "caf.cern.ch" not in fileLoc:
newfile.setLocation("caf.cern.ch")
# else:
# logging.debug("File already associated to %s" %fileLoc)
LOCK.release()
if len(newfile["runs"]):
val = 0
for run in newfile["runs"]:
if run.run < int(startRun):
val = 1
break
if not val:
listFile = fileInFileset.execute(filesetToProcess.id)
if {"fileid": newfile["id"]} not in listFile:
filesetToProcess.addFile(newfile)
filesetToProcess.setLastUpdate(int(newFilesList["end_time"]) + 1)
filesetToProcess.commit()
logging.debug("new file created/loaded added by T0AST...")
except Exception, e:
logging.debug("Error when adding new files in T0AST...")
logging.debug(e)
logging.debug(format_exc())
LOCK.release()
def __call__(self, filesetToProcess):
"""
The algorithm itself
"""
# Get configuration
initObj = WMInit()
initObj.setLogging()
initObj.setDatabaseConnection(os.getenv("DATABASE"), \
os.getenv('DIALECT'), os.getenv("DBSOCK"))
myThread = threading.currentThread()
daofactory = DAOFactory(package = "WMCore.WMBS" , \
logger = myThread.logger, \
dbinterface = myThread.dbi)
locationNew = daofactory(classname = "Locations.New")
getFileLoc = daofactory(classname = "Files.GetLocation")
fileInFileset = daofactory(classname = "Files.InFileset")
logging.debug("DBSFeeder is processing %s" % \
filesetToProcess.name)
logging.debug("the filesetBase name %s" \
% (filesetToProcess.name).split(":")[0])
LASTIME = filesetToProcess.lastUpdate
# Get the start Run if asked
startRun = (filesetToProcess.name).split(":")[3]
# get list of files
tries = 1
while True:
try:
blocks = self.dbsReader.getFiles(\
(filesetToProcess.name).split(":")[0])
now = time.time()
logging.debug("DBS queries done ...")
break
except DBSReaderError as ex:
logging.error("DBS error: %s, cannot get files for %s" % \
(str(ex), filesetToProcess.name))
# Close fileset
filesetToProcess.markOpen(False)
return
# connection error, retry
except DbsConnectionError as ex:
logging.error("Unable to connect to DBS, retrying: " + \
str(ex))
if tries > self.connectionAttempts: #too many errors - bail out
return
tries = tries + 1
# check for empty datasets
if blocks == {}:
logging.debug("DBS: Empty blocks - %s" %filesetToProcess.name)
return filesetToProcess
# get all file blocks
blockList = blocks.keys()
# process all file blocks
for fileBlock in blockList:
seList = blocks[fileBlock]['StorageElements']
# add files for non blocked SE
if seList is None or seList == []:
logging.info("fileblock %s - no SE's associated" % \
fileBlock)
continue
else:
for loc in seList:
locationNew.execute(siteName = loc, seName = loc)
for files in blocks[fileBlock]['Files']:
if startRun != 'None':
if len(files['LumiList']):
for lumi in files['LumiList']:
if int(startRun) <= int(lumi['RunNumber' ]):
newfile = File(files['LogicalFileName'], \
size=files['FileSize'], events=files\
['NumberOfEvents'])
LOCK.acquire()
if newfile.exists() == False :
newfile.create()
filesetToProcess.addFile(newfile)
filesetToProcess.setLastUpdate(\
int(time.time()))
filesetToProcess.commit()
runSet = set()
runSet.add(Run( lumi\
['RunNumber' ], *[lumi['LumiSectionNumber']] ))
newfile.addRunSet(runSet)
else:
newfile.loadData()
listFile = fileInFileset.execute\
(filesetToProcess.id)
if {'fileid': newfile[\
'id']} not in listFile:
filesetToProcess.addFile(newfile)
filesetToProcess.setLastUpdate\
(int(time.time()))
filesetToProcess.commit()
val = 0
for run in newfile['runs']:
if lumi['RunNumber' ] == run.run:
val = 1
break
if not val:
runSet = set()
runSet.add(Run(\
lumi['RunNumber' ], *[lumi['LumiSectionNumber']]))
newfile.addRunSet(runSet)
fileLoc = getFileLoc.execute(\
file = files['LogicalFileName'])
if fileLoc:
for loc in seList:
if loc not in fileLoc:
newfile.setLocation(\
loc)
else:
newfile.setLocation(seList)
LOCK.release()
else:
# Assume parents and LumiSection aren't asked
newfile = File(files['LogicalFileName'], \
size=files['FileSize'], events=files['NumberOfEvents'])
LOCK.acquire()
if newfile.exists() == False :
newfile.create()
# Update fileset last update parameter
filesetToProcess.addFile(newfile)
logging.debug("new file created and added by DBS")
filesetToProcess.setLastUpdate(int(time.time()))
filesetToProcess.commit()
else:
newfile.loadData()
listFile = fileInFileset.execute(filesetToProcess.id)
if {'fileid': newfile['id']} not in listFile:
filesetToProcess.addFile(newfile)
logging.debug("new file loaded and added by DBS")
filesetToProcess.setLastUpdate(int(time.time()))
filesetToProcess.commit()
fileLoc = getFileLoc.execute(\
file = files['LogicalFileName'])
if fileLoc:
for loc in seList:
if loc not in fileLoc:
newfile.setLocation(loc)
else:
newfile.setLocation(seList)
LOCK.release()
filesetToProcess.load()
LASTIME = filesetToProcess.lastUpdate
# For re-opned fileset or empty, try until the purge time
if (int(now)/3600 - LASTIME/3600) > self.reopenTime:
filesetToProcess.setLastUpdate(int(time.time()))
filesetToProcess.commit()
if (int(now)/3600 - LASTIME/3600) > self.purgeTime:
filesetToProcess.markOpen(False)
logging.debug("Purge Done...")
filesetToProcess.commit()
logging.debug("DBS feeder work done...")
