def _queryAndCompareWithDBS(self, pileupDict, defaultArguments, dbsUrl):
"""
pileupDict is a Python dictionary containing particular pileup
configuration information. Query DBS on given dataset contained
now in both input defaultArguments as well as in the pileupDict
and compare values.
"""
args = {}
# this should have been set in CMSSWStepHelper along with
# the pileup configuration
args["url"] = dbsUrl
args["version"] = "DBS_2_0_9"
args["mode"] = "GET"
dbsApi = DbsApi(args)
inputArgs = defaultArguments["PileupConfig"]
self.assertEqual(len(inputArgs), len(pileupDict),
"Number of pileup types different.")
for pileupType in inputArgs:
m = ("pileup type '%s' not in PileupFetcher-produced pileup "
"configuration: '%s'" % (pileupType, pileupDict))
self.failUnless(pileupType in pileupDict, m)
# now query DBS for compare actual results on files lists for each
# pileup type and dataset and location (storage element names)
# pileupDict is saved in the file and now comparing items of this
# configuration with actual DBS results, the structure of pileupDict:
# {"pileupTypeA": {"BlockA": {"FileList": [], "StorageElementNames": []},
# "BlockB": {"FileList": [], "StorageElementName": []}, ....}
for pileupType, datasets in inputArgs.items():
# this is from the pileup configuration produced by PileupFetcher
blockDict = pileupDict[pileupType]
m = "Number of datasets for pileup type '%s' is not equal." % pileupType
self.assertEqual(len(blockDict), len(datasets), m)
for dataset in datasets:
dbsFileBlocks = dbsApi.listBlocks(dataset = dataset)
fileList = [] # list of files in the block (dbsFile["LogicalFileName"])
storageElemNames = [] # list of StorageElementName
for dbsFileBlock in dbsFileBlocks:
blockName = dbsFileBlock["Name"]
# each DBS block has a list under 'StorageElementList', iterate over
for storElem in dbsFileBlock["StorageElementList"]:
storageElemNames.append(storElem["Name"])
# now get list of files in the block
dbsFiles = dbsApi.listFiles(blockName = blockName)
for dbsFile in dbsFiles:
fileList.append(dbsFile["LogicalFileName"])
# now compare the sets:
m = ("StorageElementNames don't agree for pileup type '%s', "
"dataset '%s' in configuration: '%s'" % (pileupType, dataset, pileupDict))
self.assertEqual(blockDict[blockName]["StorageElementNames"], storageElemNames, m)
m = ("FileList don't agree for pileup type '%s', dataset '%s' "
" in configuration: '%s'" % (pileupType, dataset, pileupDict))
self.assertEqual(blockDict[blockName]["FileList"], fileList)
args = {}
args['url'] = dstURL
args['mode']='POST'
api = DbsApi(args)
try:
print "\n\nListing Datasets "
paths = api.listDatasetPaths()
ranIndex = random.randint(1,len(paths))
#myDataset = paths[ranIndex]
myDataset = "/Wjets-sherpa/Summer08_IDEAL_V12_v1/GEN-SIM-RAW"
print "Selected dataset is %s", myDataset
print "\nListing Blocks "
blocks = api.listBlocks(myDataset)
#print blocks
print "\n Listing Files "
files = api.listFiles(path = myDataset, retriveList=["all"])
#print files
#for path in paths:
# print path
except DbsApiException, ex:
print "Caught API Exception %s: %s " % (ex.getClassName(), ex.getErrorMessage() )
if ex.getErrorCode() not in (None, ""):
print "DBS Exception Error Code: ", ex.getErrorCode()
#!/usr/bin/env python
#
# API Unit tests for the DBS JavaServer.
import sys
from DBSAPI.dbsApi import DbsApi
from DBSAPI.dbsException import *
from DBSAPI.dbsApiException import *
from DBSAPI.dbsOptions import DbsOptionParser
try:
optManager = DbsOptionParser()
(opts,args) = optManager.getOpt()
api = DbsApi(opts.__dict__)
for block in api.listBlocks("/test_primary_anzar_001/SIM/TestProcessedDS001"):
#for block in api.listBlocks("/TestPrimary1167862926.47/SIM1167862926.47/TestProcessed1167862926.47", "/*hahah#12345"):
#for block in api.listBlocks("/TestPrimary1167862926.47/SIM1167862926.47/TestProcessed1167862926.47", "/this/*"):
#for block in api.listBlocks("/TestPrimary1167862926.47/SIM1167862926.47/TestProcessed1167862926.47", "/this/ff*"):
#for block in api.listBlocks("/TestPrimary1167862926.47/SIM1167862926.47/TestProcessed1167862926.47", "/this/hahah#12345"):
#for block in api.listBlocks("/TestPrimary1167862926.47/SIM1167862926.47/TestProcessed1167862926.47", "/this/hahah#12345"):
print " %s" % block
except DbsApiException, ex:
print "Caught API Exception %s: %s " % (ex.getClassName(), ex.getErrorMessage() )
if ex.getErrorCode() not in (None, ""):
print "DBS Exception Error Code: ", ex.getErrorCode()
if blockFile not in blockFiles:
if not badFiles.has_key(blockName):
badFiles[blockName] = []
badFiles[blockName].append(blockFile)
#sys.exit(0)
psetInstance = DbsQueryableParameterSet(Hash = "GIBBERISH")
for newBlockName in badFiles.keys():
seName = blockLocation[newBlockName]
(datasetPath, junk) = newBlockName.split("#", 1)
dbsApi.insertBlock(datasetPath, newBlockName, storage_element_list = [seName])
blockRef = dbsApi.listBlocks(dataset = datasetPath, block_name = newBlockName)[0]
print blockRef
newFiles = []
for newFileLFN in badFiles[newBlockName]:
localFile = DBSBufferFile(lfn = newFileLFN)
localFile.load(parentage = 1)
(primaryDS, procDS, tier) = datasetPath[1:].split("/", 3)
primary = DbsPrimaryDataset(Name = primaryDS, Type = "mc")
algo = DbsAlgorithm(ExecutableName = localFile["appName"],
ApplicationVersion = localFile["appVer"],
ApplicationFamily = localFile["appFam"],
ParameterSetID = psetInstance)
processed = DbsProcessedDataset(PrimaryDataset = primary,
AlgoList = [algo],
apiSrc = DbsApi(opts.__dict__)
else :
apiSrc = DbsApi(argsSrc)
if(argsTar['url'] == "LOCALDBS" ):
apiTar = DbsApi(opts.__dict__)
else:
apiTar = DbsApi(argsTar)
path = sys.argv[3]
print path
#name = argsSrc['url'].replace('/','_') + "_" + argsTar['url'].replace('/', '_') + path.replace('/', '_')
name = argsSrc['url'].replace('/','_').replace(':', '_') + "_" + argsTar['url'].replace('/', '_').replace(':', '_')
blocks = apiSrc.listBlocks(path)
if ((op == "both") | (op == "get")) :
#Fetch the dataset contents and save them in a file
for i in blocks:
blockName = i['Name']
fileName = blockName.replace('/', '_').replace('#', '_') + ".xml"
print "Fetching information for Block %s " % blockName
xmlinput = apiSrc.listDatasetContents(path, blockName)
f = open(name + fileName, "w");
f.write(xmlinput)
f.close()
print "Dataset information fetched from " + argsSrc['url'] + " in XML format is saved in " + name + fileName
if ((op == "both") | (op == "set")) :
#Insert the saved contents into another DBS instance
for i in blocks:
print " # files already published: ",nPublishedFiles
print " # new files: ",nFiles
print "#################################################################################################"
#--------------------
# loop over files
count = 0
while count < len(files):
# ------------------
# insert block
print "inserting block..."
blockName = dbsApi.insertBlock(datasetPath, None ,
storage_element_list = [seName])
block = dbsApi.listBlocks(datasetPath,
block_name = blockName,
storage_element_name = seName)[0]
print "...block inserted (name: ",blockName,")"
#--------------------
# loop over files
start = count
stop = min(count+blockSize,len(files))
_files = files[start:stop]
dbsFiles = []
print "preparing files for 1 block..."
for file in _files:
badFiles[blockName] = []
badFiles[blockName].append(blockFile)
#sys.exit(0)
psetInstance = DbsQueryableParameterSet(Hash="GIBBERISH")
for newBlockName in badFiles.keys():
seName = blockLocation[newBlockName]
(datasetPath, junk) = newBlockName.split("#", 1)
dbsApi.insertBlock(datasetPath,
newBlockName,
storage_element_list=[seName])
blockRef = dbsApi.listBlocks(dataset=datasetPath,
block_name=newBlockName)[0]
print blockRef
newFiles = []
for newFileLFN in badFiles[newBlockName]:
localFile = DBSBufferFile(lfn=newFileLFN)
localFile.load(parentage=1)
(primaryDS, procDS, tier) = datasetPath[1:].split("/", 3)
primary = DbsPrimaryDataset(Name=primaryDS, Type="mc")
algo = DbsAlgorithm(ExecutableName=localFile["appName"],
ApplicationVersion=localFile["appVer"],
ApplicationFamily=localFile["appFam"],
ParameterSetID=psetInstance)
processed = DbsProcessedDataset(PrimaryDataset=primary,
AlgoList=[algo],
if myopt == "--dataset" :
datasetpath = myarg
if datasetpath == None:
print "Please specify --dataset"
sys.exit(2)
se = 'srm.cern.ch'
try:
outputname = datasetpath.replace('/','_')[1:]+'.xml'
outputfile = file(outputname.replace('PreCSA08','CSA08'),'w')
output = ""
api = DbsApi(dbsargs)
try:
blocks = api.listBlocks(dataset=datasetpath,block_name="*",storage_element_name=se);
except:
blocks = []
for block in blocks:
output += api.listDatasetContents(datasetpath,block['Name'])
newoutput = output.replace('PreCSA08','CSA08')
outputfile.write(newoutput)
outputfile.close()
except DbsApiException, ex:
print "Caught API Exception %s: %s " % (ex.getClassName(),
ex.getErrorMessage() )
if ex.getErrorCode() not in (None, ""):
print "DBS Exception Error Code: ", ex.getErrorCode()
def _queryAndCompareWithDBS(self, pileupDict, defaultArguments, dbsUrl):
"""
pileupDict is a Python dictionary containing particular pileup
configuration information. Query DBS on given dataset contained
now in both input defaultArguments as well as in the pileupDict
and compare values.
"""
args = {}
# this should have been set in CMSSWStepHelper along with
# the pileup configuration
args["url"] = dbsUrl
args["version"] = "DBS_2_0_9"
args["mode"] = "GET"
dbsApi = DbsApi(args)
inputArgs = defaultArguments["PileupConfig"]
self.assertEqual(len(inputArgs), len(pileupDict),
"Number of pileup types different.")
for pileupType in inputArgs:
m = ("pileup type '%s' not in PileupFetcher-produced pileup "
"configuration: '%s'" % (pileupType, pileupDict))
self.failUnless(pileupType in pileupDict, m)
# now query DBS for compare actual results on files lists for each
# pileup type and dataset and location (storage element names)
# pileupDict is saved in the file and now comparing items of this
# configuration with actual DBS results, the structure of pileupDict:
# {"pileupTypeA": {"BlockA": {"FileList": [], "StorageElementNames": []},
# "BlockB": {"FileList": [], "StorageElementName": []}, ....}
for pileupType, datasets in inputArgs.items():
# this is from the pileup configuration produced by PileupFetcher
blockDict = pileupDict[pileupType]
m = "Number of datasets for pileup type '%s' is not equal." % pileupType
self.assertEqual(len(blockDict), len(datasets), m)
for dataset in datasets:
dbsFileBlocks = dbsApi.listBlocks(dataset=dataset)
fileList = [
] # list of files in the block (dbsFile["LogicalFileName"])
storageElemNames = [] # list of StorageElementName
for dbsFileBlock in dbsFileBlocks:
blockName = dbsFileBlock["Name"]
# each DBS block has a list under 'StorageElementList', iterate over
for storElem in dbsFileBlock["StorageElementList"]:
storageElemNames.append(storElem["Name"])
# now get list of files in the block
dbsFiles = dbsApi.listFiles(blockName=blockName)
for dbsFile in dbsFiles:
fileList.append(dbsFile["LogicalFileName"])
# now compare the sets:
m = ("StorageElementNames don't agree for pileup type '%s', "
"dataset '%s' in configuration: '%s'" %
(pileupType, dataset, pileupDict))
self.assertEqual(blockDict[blockName]["StorageElementNames"],
storageElemNames, m)
m = ("FileList don't agree for pileup type '%s', dataset '%s' "
" in configuration: '%s'" %
(pileupType, dataset, pileupDict))
self.assertEqual(blockDict[blockName]["FileList"], fileList)
class BlcokConsitencyTest(object):
def __init__(self):
paConfig = loadProdAgentConfiguration()
t0astDBConfig = paConfig.getConfig("Tier0DB")
self.t0astDBConn = Tier0DB.Tier0DB(t0astDBConfig, manageGlobal = True)
self.t0astDBConn.connect()
localDBSConfig = paConfig.getConfig("LocalDBS")
globalDBSConfig = paConfig.getConfig("GlobalDBSDLS")
# phedexConfig = paConfig.getConfig("PhEDExConfig")
self.localDBSUrl = localDBSConfig["DBSURL"]
self.globalDBSUrl = globalDBSConfig["DBSURL"]
# self.phedexDSUrl = phedexConfig["DataServiceURL"]
self.localDbsApi = DbsApi({'url':localDBSConfig["DBSURL"]})
self.globalDbsApi = DbsApi({'url':globalDBSConfig["DBSURL"]})
def listMissingBlocksFromDBS(self, dbsApi, blockStatus, migrateStatus):
blockList = ListBlock.listBlocksByStatus(self.t0astDBConn, blockStatus)
missingBlockList = []
for block in blockList:
dbsBlock = dbsApi.listBlocks(block_name=block["BLOCK_NAME"])
if dbsBlock == []:
missingBlockList.append(block)
return missingBlockList
def listAllBlocksFromDBS(self, dbsApi, blockStatus, migrateStatus):
blockList = ListBlock.listBlocksByStatus(self.t0astDBConn, blockStatus, migrateStatus)
dbsBlockList = []
for block in blockList:
dbsBlock = dbsApi.listBlocks(block_name=block["BLOCK_NAME"])
if dbsBlock != []:
dbsBlockList.append(block)
return dbsBlockList
def printMissingBlockName(self, location, status, migrateStatus = None, verbose = 1):
if location == "local":
dbsApi = self.localDbsApi
dbsUrl = self.localDBSUrl
elif location == "global":
dbsApi = self.globalDbsApi
dbsUrl = self.globalDBSUrl
verboseLevel = int(verbose)
print "For the all %s, %s block -" % (status, migrateStatus)
if verboseLevel > 2:
allBlockList = self.listAllBlocksFromDBS(dbsApi, status, migrateStatus)
print "All t0ast blocks in DBS : %s number: %s" % (dbsUrl, len(allBlockList))
self.printBlocksAndFiles(location, allBlockList, "all", verboseLevel)
blockList = self.listMissingBlocksFromDBS(dbsApi, status, migrateStatus)
if blockList == []:
print "There is no missing blocks in DBS : %s" % dbsUrl
else:
print "List of missing blocks from DBS : %s number: %s" % (dbsUrl, len(blockList))
self.printBlocksAndFiles(location, blockList, "missing", verboseLevel)
print "the number of missing block: %s" % len(blockList)
def printBlocksAndFiles(self, location, blockList, type = "missing", verbose = 1):
"""
print blocks and files: verbose = 1 only blocks, verbose > 1 blocks and files
"""
for block in blockList:
print "\n"
print "#################################################################"
print "Block ID: %s Name: %s" % (block["BLOCK_ID"], block["BLOCK_NAME"])
if verbose < 2:
continue
if location == "global":
print "\n"
print "List all files from %s block from Global DBS: Get info from local DBS " % type
for file in self.localDbsApi.listFiles(blockName=block["BLOCK_NAME"]):
print "parent list: %s" % file["ParentList"]
print "file LFN: %s" % file["LogicalFileName"]
print "\n"
fileIDs = ListFiles.listFileIDsByBlockID(self.t0astDBConn, block["BLOCK_ID"])
print "List all files from %s block from Global DBS: Get info from T0AST " % type
print "====================================================="
for fileID in fileIDs:
wmbsFile = WMBSFile(id = fileID)
wmbsFile.load()
print "--------------------------------------"
print "Info from: T0AST"
print "file LFN: %s" % wmbsFile["lfn"]
print ""
file = T0ASTFile(wmbsFile)
file.datasetPathID = \
ListDatasets.listDatasetIDForWMBSFile(self.t0astDBConn, wmbsFile["id"])
datasetNames = \
ListDatasets.listDatasetNamesForWMBSFile(self.t0astDBConn, wmbsFile["id"])
file["PRIMARY_DATASET"] = datasetNames["PRIMARY"]
file["PROCESSED_DATASET"] = datasetNames["PROCESSED"]
file["DATA_TIER"] = datasetNames["TIER"]
if file["DATA_TIER"] == "RECO":
t0ParentFileList = file.getParentList(type="file")
for wmbsFile in t0ParentFileList:
t0File = T0ASTFile(wmbsFile)
t0File["BLOCK_ID"] = ListBlock.getBlockIDByFileID(self.t0astDBConn, wmbsFile["id"])
print "Block ID: %s : Parent File: %s" % (t0File["BLOCK_ID"], t0File["LFN"])
if t0File["BLOCK_ID"] != None:
blockInfo = ListBlock.getBlockInfoByID(self.t0astDBConn, t0File["BLOCK_ID"])
print "Block Name: %s \nStatus: %s" % (blockInfo["BLOCK_NAME"], blockInfo["STATUS"])
if blockInfo["STATUS"] == "InFlight" or blockInfo["MIGRATE_STATUS"] == "Migrated":
dbsBlock = self.localDbsApi.listBlocks(block_name=blockInfo["BLOCK_NAME"])
if dbsBlock == []:
print "It doesn't exist in Local dbs: Something wrong"
else:
print "Block: %s exist in Local DBS" % blockInfo["BLOCK_NAME"]
try:
for file in self.localDbsApi.listFiles(patternLFN=t0File["LFN"]):
print "File: %s exist in Local DBS" % file["LogicalFileName"]
except:
print "File doesn't exist in Local DBS"
if blockInfo["MIGRATE_STATUS"] == "Migrated":
dbsBlock = self.globalDbsApi.listBlocks(block_name=blockInfo["BLOCK_NAME"])
if dbsBlock == []:
print "It doesn't exist in Global dbs: Something wrong"
else:
print "Block: %s exist in Global DBS" % blockInfo["BLOCK_NAME"]
try:
for file in self.globalDbsApi.listFiles(patternLFN=t0File["LFN"]):
print "File: %s exist in Global DBS" % file["LogicalFileName"]
except:
print "File doesn't exist in Global DBS"
print "\n"
print "Info from Local DBS: List all parent files from %s block:" % type
try:
for pfile in self.localDbsApi.listFileParents(file["LFN"]):
print "Parent Block: %s" % pfile["Block"]["NAME"]
print "Parent File: %s" % pfile["LogicalFileName"]
print "Info from Global DBS: parent block for %s block:" % type
blockList = self.globalDbsApi.listBlocks(block_name=block["BLOCK_NAME"])
if blockList == []:
print "Global DBS doen't have block %s "% pfile["Block"]["NAME"]
else:
for dbsBlock in blockList:
print "Global DBS Parent block %s exsist" % dbsBlock["NAME"]
except Exception, ex:
print "No parents file found in Local DBS "
print "====================================================="
else :
apiSrc = DbsApi(argsSrc)
if(argsTar['url'] == "LOCALDBS" ):
apiTar = DbsApi(opts.__dict__)
else:
apiTar = DbsApi(argsTar)
path = sys.argv[3]
print path
#name = argsSrc['url'].replace('/','_') + "_" + argsTar['url'].replace('/', '_') + path.replace('/', '_')
#name = argsSrc['url'].replace('/','_').replace(':', '_') + "_" + argsTar['url'].replace('/', '_').replace(':', '_')
name = "Transfer_"
blocks = apiSrc.listBlocks(path)
if ((op == "both") | (op == "get")) :
#Fetch the dataset contents and save them in a file
for i in blocks:
blockName = i['Name']
fileName = blockName.replace('/', '_').replace('#', '_') + ".xml"
if os.path.exists(name + fileName):
print "WARNNING The XML file " + name + fileName + " exists already and will be used. The information will not be fetched again"
else:
print "Fetching information for Block %s " % blockName
xmlinput = apiSrc.listDatasetContents(path, blockName)
f = open(name + fileName, "w");
f.write(xmlinput)
f.close()
print "Dataset information fetched from " + argsSrc['url'] + " in XML format is saved in " + name + fileName
except getopt.GetoptError:
print "Please specify --dataset"
sys.exit(2)
for myopt, myarg in myopts :
if myopt == "--dataset" :
datasetpath = myarg
if datasetpath == None:
print "Please specify --dataset"
sys.exit(2)
cern_se = 'srm.cern.ch'
try:
api = DbsApi(dbsargs)
try:
blocks = api.listBlocks(dataset=datasetpath,block_name="*",storage_element_name="*");
except:
blocks = []
for block in blocks:
for se in block['StorageElementList'] :
if se['Name'] != cern_se :
api.deleteReplicaFromBlock(block['Name'],se['Name'])
except DbsApiException, ex:
print "Caught API Exception %s: %s " % (ex.getClassName(),
ex.getErrorMessage() )
if ex.getErrorCode() not in (None, ""):
print "DBS Exception Error Code: ", ex.getErrorCode()
class DBSWriter:
"""
_DBSWriter_
General API for writing data to DBS
"""
def __init__(self, url, **contact):
args = {"url": url, "level": 'ERROR'}
args.update(contact)
try:
self.dbs = DbsApi(args)
self.args = args
self.version = args.get('version', None)
self.globalDBSUrl = args.get('globalDBSUrl', None)
self.globalVersion = args.get('globalVersion', None)
if self.globalDBSUrl:
globalArgs = {'url': url, 'level': 'ERROR'}
globalArgs.update(contact)
self.globalDBS = DbsApi(globalArgs)
except DbsException as ex:
msg = "Error in DBSWriterError with DbsApi\n"
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
self.reader = DBSReader(**args)
def createDatasets(self, workflowSpec):
"""
_createDatasets_
Create All the output datasets found in the workflow spec instance
provided
"""
try:
workflowSpec.payload.operate(
_CreateDatasetOperator(self.dbs, workflowSpec))
except DbsException as ex:
msg = "Error in DBSWriter.createDatasets\n"
msg += "For Workflow: %s\n" % workflowSpec.workflowName()
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
return
def insertFilesForDBSBuffer(self,
files,
procDataset,
algos,
jobType="NotMerge",
insertDetectorData=False,
maxFiles=100,
maxSize=99999999,
timeOut=None,
fileCommitLength=5):
"""
_insertFiles_
list of files inserted in DBS
"""
#TODO: Whats the purpose of insertDetectorData
if len(files) < 1:
return
affectedBlocks = []
insertFiles = []
addedRuns = []
seName = None
#Get the algos in insertable form
# logging.error("About to input algos")
# logging.error(algos)
ialgos = [
DBSWriterObjects.createAlgorithmForInsert(dict(algo))
for algo in algos
]
#print ialgos
for outFile in files:
# //
# // Convert each file into a DBS File object
#//
lumiList = []
#Somehing similar should be the real deal when multiple runs/lumi could be returned from wmbs file
for runlumiinfo in outFile.getRuns():
lrun = long(runlumiinfo.run)
run = DbsRun(
RunNumber=lrun,
NumberOfEvents=0,
NumberOfLumiSections=0,
TotalLuminosity=0,
StoreNumber=0,
StartOfRun=0,
EndOfRun=0,
)
#Only added if not added by another file in this loop, why waste a call to DBS
if lrun not in addedRuns:
self.dbs.insertRun(run)
addedRuns.append(
lrun) #save it so we do not try to add it again to DBS
logging.debug("run %s added to DBS " % str(lrun))
for alsn in runlumiinfo:
lumi = DbsLumiSection(
LumiSectionNumber=long(alsn),
StartEventNumber=0,
EndEventNumber=0,
LumiStartTime=0,
LumiEndTime=0,
RunNumber=lrun,
)
lumiList.append(lumi)
logging.debug("lumi list created for the file")
dbsfile = DbsFile(
#Checksum = str(outFile['cksum']),
NumberOfEvents=outFile['events'],
LogicalFileName=outFile['lfn'],
FileSize=int(outFile['size']),
Status="VALID",
ValidationStatus='VALID',
FileType='EDM',
Dataset=procDataset,
TierList=DBSWriterObjects.makeTierList(
procDataset['Path'].split('/')[3]),
AlgoList=ialgos,
LumiList=lumiList,
ParentList=outFile.getParentLFNs(),
#BranchHash = outFile['BranchHash'],
)
#Set checksums by hand
#dbsfile['Checksum'] = 0 #Set a default?
for entry in outFile['checksums'].keys():
#This should be a dictionary with a cktype key and cksum value
if entry.lower() == 'cksum':
dbsfile['Checksum'] = str(outFile['checksums'][entry])
elif entry.lower() == 'adler32':
dbsfile['Adler32'] = str(outFile['checksums'][entry])
elif entry.lower() == 'md5':
dbsfile['Md5'] = str(outFile['checksums'][entry])
#This check comes from ProdAgent, not sure if its required
if len(outFile["locations"]) > 0:
seName = list(outFile["locations"])[0]
logging.debug("SEname associated to file is: %s" % seName)
else:
msg = "Error in DBSWriter.insertFiles\n"
msg += "No SEname associated to file"
#print "FAKING seName for now"
#seName="cmssrm.fnal.gov"
raise DBSWriterError(msg)
insertFiles.append(dbsfile)
# //Processing Jobs:
# // Insert the lists of sorted files into the appropriate
#// fileblocks
sumSize = 0
sumFiles = 0
tmpFiles = []
blockList = []
#First, get the block. See if the block already exists
try:
fileBlock = DBSWriterObjects.getDBSFileBlock(
self.dbs, procDataset, seName)
fileBlock['files'] = []
#if not fileBlock in affectedBlocks:
# affectedBlocks.append(fileBlock)
except DbsException as ex:
msg = "Error in DBSWriter.insertFilesForDBSBuffer\n"
msg += "Cannot retrieve FileBlock for dataset:\n"
msg += " %s\n" % procDataset['Path']
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
filesToCommit = []
for file in insertFiles:
# First see if the block is full
if self.manageFileBlock(fileBlock=fileBlock,
maxFiles=maxFiles,
maxSize=maxSize,
timeOut=timeOut,
algos=ialgos,
filesToCommit=filesToCommit,
procDataset=procDataset):
fileBlock['OpenForWriting'] = 0
if not fileBlock in affectedBlocks:
affectedBlocks.append(fileBlock)
# Then we need a new block
try:
fileBlock = DBSWriterObjects.getDBSFileBlock(
self.dbs, procDataset, seName)
fileBlock['files'] = []
except DbsException as ex:
msg = "Error in DBSWriter.insertFilesForDBSBuffer\n"
msg += "Cannot retrieve FileBlock for dataset:\n"
msg += " %s\n" % procDataset['Path']
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
fileBlock['files'].append(file['LogicalFileName'])
filesToCommit.append(file)
if len(filesToCommit) >= fileCommitLength:
# Only commit the files if there are more of them then the maximum length
try:
self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
filesToCommit = []
logging.debug("Inserted files: %s to FileBlock: %s" \
% ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
msg += " %s\n" % (
[x['LogicalFileName'] for x in insertFiles], )
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
if len(filesToCommit) > 0:
try:
self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
filesToCommit = []
logging.debug("Inserted files: %s to FileBlock: %s" \
% ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
msg += " %s\n" % ([x['LogicalFileName']
for x in insertFiles], )
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
if not fileBlock in affectedBlocks:
affectedBlocks.append(fileBlock)
## Do bulk inserts now for DBS
#filesToCommit = []
#count = 0
#count2 = 0
#for file in insertFiles:
# count += 1
# #Try and close the box
# logging.error("Should have a file")
# logging.error(len(filesToCommit))
# count2 += len(filesToCommit)
# if self.manageFileBlock(fileBlock = fileBlock, maxFiles = maxFiles,
# maxSize = maxSize, timeOut = timeOut, algos = ialgos,
# filesToCommit = filesToCommit, procDataset = procDataset):
# fileBlock['OpenForWriting'] = '0'
# if not fileBlock in affectedBlocks:
# affectedBlocks.append(fileBlock)
#
#
#
# # Then we need a new block
# try:
# fileBlock = DBSWriterObjects.getDBSFileBlock(
# self.dbs,
# procDataset,
# seName)
# fileBlock['files'] = []
# except DbsException, ex:
# msg = "Error in DBSWriter.insertFilesForDBSBuffer\n"
# msg += "Cannot retrieve FileBlock for dataset:\n"
# msg += " %s\n" % procDataset['Path']
# msg += "%s\n" % formatEx(ex)
# raise DBSWriterError(msg)
# #At this point, we should commit the block as is
# fileBlock['files'].append(file['LogicalFileName'])
# if jobType == "MergeSpecial":
# for file in fileList:
# file['Block'] = fileBlock
# msg="calling: self.dbs.insertMergedFile(%s, %s)" % (str(file['ParentList']),str(file))
# logging.debug(msg)
# try:
# #
# #
# # NOTE To Anzar From Anzar (File cloning as in DBS API can be done here and then I can use Bulk insert on Merged files as well)
# self.dbs.insertMergedFile(file['ParentList'],
# file)
#
# except DbsException, ex:
# msg = "Error in DBSWriter.insertFiles\n"
# msg += "Cannot insert merged file:\n"
# msg += " %s\n" % file['LogicalFileName']
# msg += "%s\n" % formatEx(ex)
# raise DBSWriterError(msg)
# logging.debug("Inserted merged file: %s to FileBlock: %s"%(file['LogicalFileName'],fileBlock['Name']))
# else:
# filesToCommit.append(file)
# if len(filesToCommit) >= fileCommitLength:
# # Only commit the files if there are more of them then the maximum length
# try:
# logging.error("About to commit %i files" %(len(filesToCommit)))
# count2 += len(filesToCommit)
# self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
# filesToCommit = []
# logging.debug("Inserted files: %s to FileBlock: %s" \
# % ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
#
# except DbsException, ex:
# msg = "Error in DBSWriter.insertFiles\n"
# msg += "Cannot insert processed files:\n"
# msg += " %s\n" % ([ x['LogicalFileName'] for x in insertFiles ],)
# msg += "%s\n" % formatEx(ex)
# raise DBSWriterError(msg)
#
#
#
#
## If we still have files to commit, commit them
#logging.error("Got to the end of the loop")
#logging.error(len(filesToCommit))
#logging.error(count2)
#if len(filesToCommit) > 0:
# try:
# logging.error("About to insert some files")
# self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
# filesToCommit = []
# logging.debug("Inserted files: %s to FileBlock: %s" \
# % ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
#
# except DbsException, ex:
# msg = "Error in DBSWriter.insertFiles\n"
# msg += "Cannot insert processed files:\n"
# msg += " %s\n" % ([ x['LogicalFileName'] for x in insertFiles ],)
# msg += "%s\n" % formatEx(ex)
# raise DBSWriterError(msg)
if not fileBlock in affectedBlocks:
affectedBlocks.append(fileBlock)
return list(affectedBlocks)
def insertFiles(self, fwkJobRep, insertDetectorData=False):
"""
_insertFiles_
Process the files in the FwkJobReport instance and insert
them into the associated datasets
A list of affected fileblock names is returned both for merged
and unmerged fileblocks. Only merged blocks will have to be managed.
#for merged file
#blocks to facilitate management of those blocks.
#This list is not populated for processing jobs since we dont really
#care about the processing job blocks.
"""
insertLists = {}
orderedHashes = []
affectedBlocks = set()
if len(fwkJobRep.files) <= 0:
msg = "Error in DBSWriter.insertFiles\n"
msg += "No files found in FrameWorkJobReport for:\n"
msg += "==> JobSpecId: %s" % fwkJobRep.jobSpecId
msg += " Workflow: %s" % fwkJobRep.workflowSpecId
raise DBSWriterError(msg)
for outFile in fwkJobRep.sortFiles():
# //
# // Convert each file into a DBS File object
#//
seName = None
if "SEName" in outFile:
if outFile['SEName']:
seName = outFile['SEName']
logging.debug("SEname associated to file is: %s" % seName)
## remove the fallback to site se-name if no SE is associated to File
## because it's likely that there is some stage out problem if there
## is no SEName associated to the file.
# if not seName:
# if fwkJobRep.siteDetails.has_key("se-name"):
# seName = fwkJobRep.siteDetails['se-name']
# seName = str(seName)
# logging.debug("site SEname: %s"%seName)
if not seName:
msg = "Error in DBSWriter.insertFiles\n"
msg += "No SEname associated to files in FrameWorkJobReport for "
# msg += "No SEname found in FrameWorkJobReport for "
msg += "==> JobSpecId: %s" % fwkJobRep.jobSpecId
msg += " Workflow: %s" % fwkJobRep.workflowSpecId
raise DBSWriterError(msg)
try:
if (insertDetectorData):
dbsFiles = DBSWriterObjects.createDBSFiles(
outFile, fwkJobRep.jobType, self.dbs)
else:
dbsFiles = DBSWriterObjects.createDBSFiles(
outFile, fwkJobRep.jobType)
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles:\n"
msg += "Error creating DbsFile instances for file:\n"
msg += "%s\n" % outFile['LFN']
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
if len(dbsFiles) <= 0:
msg = "No DbsFile instances created. Not enough info in the FrameWorkJobReport for"
msg += "==> JobSpecId: %s" % fwkJobRep.jobSpecId
msg += " Workflow: %s" % fwkJobRep.workflowSpecId
raise DBSWriterError(msg)
for f in dbsFiles:
datasetName = makeDBSDSName(f)
hashName = "%s-%s" % (seName, datasetName)
if hashName not in insertLists:
insertLists[hashName] = _InsertFileList(
seName, datasetName)
insertLists[hashName].append(f)
if not orderedHashes.count(hashName):
orderedHashes.append(hashName)
# //Processing Jobs:
# // Insert the lists of sorted files into the appropriate
#// fileblocks
for hash in orderedHashes:
fileList = insertLists[hash]
procDataset = fileList[0]['Dataset']
try:
fileBlock = DBSWriterObjects.getDBSFileBlock(
self.dbs, procDataset, fileList.seName)
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot retrieve FileBlock for dataset:\n"
msg += " %s\n" % procDataset
msg += "In Storage Element:\n %s\n" % fileList.seName
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
if fwkJobRep.jobType == "Merge":
# //
# // Merge files
#//
for mergedFile in fileList:
mergedFile['Block'] = fileBlock
affectedBlocks.add(fileBlock['Name'])
msg = "calling: self.dbs.insertMergedFile(%s, %s)" % (str(
mergedFile['ParentList']), str(mergedFile))
logging.debug(msg)
try:
self.dbs.insertMergedFile(mergedFile['ParentList'],
mergedFile)
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert merged file:\n"
msg += " %s\n" % mergedFile['LogicalFileName']
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
logging.debug(
"Inserted merged file: %s to FileBlock: %s" %
(mergedFile['LogicalFileName'], fileBlock['Name']))
else:
# //
# // Processing files
#//
affectedBlocks.add(fileBlock['Name'])
msg = "calling: self.dbs.insertFiles(%s, %s, %s)" % (
str(procDataset), str(list(fileList)), str(fileBlock))
logging.debug(msg)
try:
self.dbs.insertFiles(procDataset, list(fileList),
fileBlock)
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
msg += " %s\n" % ([x['LogicalFileName']
for x in fileList], )
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
logging.debug("Inserted files: %s to FileBlock: %s" %
(([x['LogicalFileName']
for x in fileList]), fileBlock['Name']))
return list(affectedBlocks)
def manageFileBlock(self,
fileBlock,
maxFiles=100,
maxSize=None,
timeOut=None,
algos=[],
filesToCommit=[],
procDataset=None):
"""
_manageFileBlock_
Check to see wether the fileblock with the provided name
is closeable based on number of files or total size.
If the block equals or exceeds wither the maxFiles or maxSize
parameters, close the block and return True, else do nothing and
return False
"""
# //
# // Check that the block exists, and is open before we close it
#//
fileblockName = fileBlock['Name']
blockInstance = self.dbs.listBlocks(block_name=fileblockName)
if len(blockInstance) > 1:
msg = "Multiple Blocks matching name: %s\n" % fileblockName
msg += "Unable to manage file block..."
raise DBSWriterError(msg)
if len(blockInstance) == 0:
msg = "Block name %s not found\n" % fileblockName
msg += "Cant manage a non-existent fileblock"
raise DBSWriterError(msg)
blockInstance = blockInstance[0]
isClosed = blockInstance.get('OpenForWriting', '1')
if isClosed != '1':
msg = "Block %s already closed" % fileblockName
logging.warning(msg)
# Now we need to commit files
if len(filesToCommit) > 0:
try:
self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
filesToCommit = []
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
raise DBSWriterError(msg)
# Attempting to migrate to global
if self.globalDBSUrl:
self.dbs.dbsMigrateBlock(
srcURL=self.args['url'],
dstURL=self.globalDBSUrl,
block_name=fileblockName,
srcVersion=self.version,
dstVersion=self.globalVersion,
)
#for algo in algos:
# self.globalDBS.insertAlgoInPD(dataset = get_path(fileblockName.split('#')[0]),
# algorithm = algo)
logging.info(
"Migrated block %s to global due to pre-closed status" %
(fileblockName))
else:
logging.error(
"Should've migrated block %s because it was already closed, but didn't"
% (fileblockName))
return True
# //
# // We have an open block, sum number of files and file sizes
#//
#fileCount = int(blockInstance.get('NumberOfFiles', 0))
fileCount = len(fileBlock['files'])
totalSize = float(blockInstance.get('BlockSize', 0))
msg = "Fileblock: %s\n ==> Size: %s Files: %s\n" % (
fileblockName, totalSize, fileCount)
logging.warning(msg)
# //
# // Test close block conditions
#//
closeBlock = False
if timeOut:
if int(time.time()) - int(blockInstance['CreationDate']) > timeOut:
closeBlock = True
msg = "Closing Block based on timeOut: %s" % fileblockName
logging.debug(msg)
if fileCount >= maxFiles:
closeBlock = True
msg = "Closing Block Based on files: %s" % fileblockName
logging.debug(msg)
if maxSize != None:
if totalSize >= maxSize:
closeBlock = True
msg = "Closing Block Based on size: %s" % fileblockName
logging.debug(msg)
if closeBlock:
# Now we need to commit files
if len(filesToCommit) > 0:
try:
self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
filesToCommit = []
#logging.debug("Inserted files: %s to FileBlock: %s" \
# % ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
#msg += " %s\n" % ([ x['LogicalFileName'] for x in insertFiles ],)
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
# //
# // Close the block
#//
self.dbs.closeBlock(
DBSWriterObjects.createDBSFileBlock(fileblockName))
if self.globalDBSUrl:
self.dbs.dbsMigrateBlock(srcURL=self.args['url'],
dstURL=self.globalDBSUrl,
block_name=fileblockName,
srcVersion=self.version,
dstVersion=self.globalVersion)
for algo in algos:
pass
#self.globalDBS.insertAlgoInPD(dataset = get_path(fileblockName.split('#')[0]),
# algorithm = algo)
logging.info("Migrated block %s to global" % (fileblockName))
else:
logging.error("Should've migrated block %s, but didn't" %
(fileblockName))
return closeBlock
def migrateDatasetBlocks(self, inputDBSUrl, datasetPath, blocks):
"""
_migrateDatasetBlocks_
Migrate the list of fileblocks provided by blocks, belonging
to the dataset specified by the dataset path to this DBS instance
from the inputDBSUrl provided
- *inputDBSUrl* : URL for connection to input DBS
- *datasetPath* : Name of dataset in input DBS (must exist in input
DBS)
- *blocks* : list of block names to be migrated (must exist
in input DBS)
"""
if len(blocks) == 0:
msg = "FileBlocks not provided.\n"
msg += "You must provide the name of at least one fileblock\n"
msg += "to be migrated"
raise DBSWriterError(msg)
# //
# // Hook onto input DBSUrl and verify that the dataset & blocks
#// exist
reader = DBSReader(inputDBSUrl)
inputBlocks = reader.listFileBlocks(datasetPath)
for block in blocks:
# //
# // Test block exists at source
#//
if block not in inputBlocks:
msg = "Block name:\n ==> %s\n" % block
msg += "Not found in input dataset:\n ==> %s\n" % datasetPath
msg += "In DBS Instance:\n ==> %s\n" % inputDBSUrl
raise DBSWriterError(msg)
# //
# // Test block does not exist in target
#//
if self.reader.blockExists(block):
# //
# // block exists
#// If block is closed dont attempt transfer
if not self.reader.blockIsOpen(block):
msg = "Block already exists in target DBS and is closed:\n"
msg += " ==> %s\n" % block
msg += "Skipping Migration of that block"
logging.warning(msg)
continue
try:
xferData = reader.dbs.listDatasetContents(datasetPath, block)
except DbsException as ex:
msg = "Error in DBSWriter.migrateDatasetBlocks\n"
msg += "Could not read content of dataset:\n ==> %s\n" % (
datasetPath, )
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
xferData = _remapBlockParentage(datasetPath, xferData)
try:
self.dbs.insertDatasetContents(xferData)
except DbsException as ex:
msg = "Error in DBSWriter.migrateDatasetBlocks\n"
msg += "Could not write content of dataset:\n ==> %s\n" % (
datasetPath, )
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
del xferData
return
def importDatasetWithExistingParents(self,
sourceDBS,
sourceDatasetPath,
targetDBS,
onlyClosed=True):
"""
_importDataset_
Import a dataset into the local scope DBS.
It complains if the parent dataset ar not there!!
- *sourceDBS* : URL for input DBS instance
- *sourceDatasetPath* : Dataset Path to be imported
- *targetDBS* : URL for DBS to have dataset imported to
"""
reader = DBSReader(sourceDBS)
inputBlocks = reader.getFileBlocksInfo(sourceDatasetPath,
onlyClosed,
locations=False)
for inputBlock in inputBlocks:
block = inputBlock['Name']
# //
# // Test block does not exist in target
#//
if self.reader.blockExists(block):
# //
# // block exists
#// If block is closed dont attempt transfer
if not str(inputBlock['OpenForWriting']) != '1':
msg = "Block already exists in target DBS and is closed:\n"
msg += " ==> %s\n" % block
msg += "Skipping Import of that block"
logging.warning(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(
inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDatasetWithExistingParents\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
logging.info("Update block locations to:")
for sename in locations:
self.dbs.addReplicaToBlock(block, sename)
logging.info(sename)
continue
try:
xferData = reader.dbs.listDatasetContents(
sourceDatasetPath, block)
except DbsException as ex:
msg = "Error in DBSWriter.importDatasetWithExistingParents\n"
msg += "Could not read content of dataset:\n ==> %s\n" % (
sourceDatasetPath, )
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
try:
self.dbs.insertDatasetContents(xferData)
except DbsException as ex:
msg = "Error in DBSWriter.importDatasetWithExistingParents\n"
msg += "Could not write content of dataset:\n ==> %s\n" % (
sourceDatasetPath, )
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
del xferData
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDatasetWithExistingParents\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
for sename in locations:
self.dbs.addReplicaToBlock(block, sename)
return
def importDataset(self,
sourceDBS,
sourceDatasetPath,
targetDBS,
onlyClosed=True):
"""
_importDataset_
Import a dataset into the local scope DBS with full parentage hirerarchy
(at least not slow because branches info is dropped)
- *sourceDBS* : URL for input DBS instance
- *sourceDatasetPath* : Dataset Path to be imported
- *targetDBS* : URL for DBS to have dataset imported to
"""
reader = DBSReader(sourceDBS)
inputBlocks = reader.getFileBlocksInfo(sourceDatasetPath,
onlyClosed,
locations=False)
blkCounter = 0
for inputBlock in inputBlocks:
block = inputBlock['Name']
# //
# // Test block does not exist in target
#//
blkCounter = blkCounter + 1
msg = "Importing block %s of %s: %s " % (blkCounter,
len(inputBlocks), block)
logging.debug(msg)
if self.reader.blockExists(block):
# //
# // block exists
#// If block is closed dont attempt transfer
if str(inputBlock['OpenForWriting']) != '1':
msg = "Block already exists in target DBS and is closed:\n"
msg += " ==> %s\n" % block
msg += "Skipping Import of that block"
logging.warning(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(
inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDataset\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
logging.info("Update block locations to:")
for sename in locations:
self.dbs.addReplicaToBlock(block, sename)
logging.info(sename)
continue
try:
self.dbs.migrateDatasetContents(sourceDBS,
targetDBS,
sourceDatasetPath,
block_name=block,
noParentsReadOnly=False)
except DbsException as ex:
msg = "Error in DBSWriter.importDataset\n"
msg += "Could not write content of dataset:\n ==> %s\n" % (
sourceDatasetPath, )
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDataset\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
for sename in locations:
self.dbs.addReplicaToBlock(block, sename)
return
def importDatasetWithoutParentage(self,
sourceDBS,
sourceDatasetPath,
targetDBS,
onlyClosed=True):
"""
_importDataset_
Import a dataset into the local scope DBS with one level parentage,
however it has severe limitation on its use due to the "ReadOnly" concept.
- *sourceDBS* : URL for input DBS instance
- *sourceDatasetPath* : Dataset Path to be imported
- *targetDBS* : URL for DBS to have dataset imported to
"""
reader = DBSReader(sourceDBS)
inputBlocks = reader.getFileBlocksInfo(sourceDatasetPath,
onlyClosed,
locations=False)
for inputBlock in inputBlocks:
block = inputBlock['Name']
# //
# // Test block does not exist in target
#//
if self.reader.blockExists(block):
# //
# // block exists
#// If block is closed dont attempt transfer
if str(inputBlock['OpenForWriting']) != '1':
msg = "Block already exists in target DBS and is closed:\n"
msg += " ==> %s\n" % block
msg += "Skipping Import of that block"
logging.warning(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(
inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDatasetWithoutParentage\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
logging.info("Update block locations to:")
for sename in locations:
self.dbs.addReplicaToBlock(block, sename)
logging.info(sename)
continue
try:
self.dbs.migrateDatasetContents(sourceDBS,
targetDBS,
sourceDatasetPath,
block_name=block,
noParentsReadOnly=True)
except DbsException as ex:
msg = "Error in DBSWriter.importDatasetWithoutParentage\n"
msg += "Could not write content of dataset:\n ==> %s\n" % (
sourceDatasetPath, )
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDatasetWithoutParentage\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
for sename in locations:
self.dbs.addReplicaToBlock(block, sename)
return
def getOutputDatasetsWithPSet(payloadNode):
"""
_getOutputDatasetsWithPSet_
Extract all the information about output datasets from the
payloadNode object provided, including the {{}} format PSet cfg
Returns a list of DatasetInfo objects including App details
from the node.
"""
result = []
for item in payloadNode._OutputDatasets:
resultEntry = DatasetInfo()
resultEntry.update(item)
resultEntry["ApplicationName"] = payloadNode.application[
'Executable']
resultEntry["ApplicationProject"] = payloadNode.application[
'Project']
resultEntry["ApplicationVersion"] = payloadNode.application[
'Version']
resultEntry["ApplicationFamily"] = item.get(
"OutputModuleName", "AppFamily")
try:
config = payloadNode.cfgInterface
psetStr = config.originalContent()
resultEntry['PSetContent'] = psetStr
except Exception as ex:
resultEntry['PSetContent'] = None
result.append(resultEntry)
return _sortDatasets(result)
#!/usr/bin/env python
#
# Revision: 1.3 $"
# Id: DBSXMLParser.java,v 1.3 2006/10/26 18:26:04 afaq Exp $"
#
# API Unit tests for the DBS JavaServer.
import sys
from DBSAPI.dbsApi import DbsApi
from DBSAPI.dbsException import *
from DBSAPI.dbsApiException import *
from DBSAPI.dbsOptions import DbsOptionParser
try:
optManager = DbsOptionParser()
(opts,args) = optManager.getOpt()
api = DbsApi(opts.__dict__)
for dataset in api.listDatasetPaths():
print "\n %s" %str(dataset)
print "Dataset parent: %s" %str(api.listPathParents)
for block in api.listBlocks(dataset):
print "block: %s" %str(block['Name'])
for parent in api.listBlockParents(block_name=block['Name']):
print "Parent: %s" % str(parent['Name'])
except DbsApiException, ex:
print "Caught API Exception %s: %s " % (ex.getClassName(), ex.getErrorMessage() )
if ex.getErrorCode() not in (None, ""):
print "DBS Exception Error Code: ", ex.getErrorCode()
optManager = DbsOptionParser()
(opts,args) = optManager.getOpt()
api = DbsApi(opts.__dict__)
datasets=[
#"/Cosmics/Commissioning08-v1/RAW",
#"/TTbar/Summer09-MC_31X_V3-v1/GEN-SIM-RAW",
"/TTbar/Summer09-MC_31X_V3-v1/GEN-SIM-RECO",
"/BeamHalo/Summer09-STARTUP31X_V7_StreamMuAlBeamHaloOverlaps-v1/ALCARECO",
"/InclusiveMu15/Summer09-MC_31X_V3_7TeV-v1/GEN-SIM-RAW",
"/Cosmics/CMSSW_3_2_7-CRAFT09_R_V4_CosmicsSeq-v1/RECO",
"/Wmunu/Summer09-MC_31X_V3_7TeV_SD_L1_L2_Mu-v1/GEN-SIM-RECO",
"/Wmunu/Summer09-MC_31X_V3_7TeV_SD_Mu9-v1/GEN-SIM-RECO"
]
#datasets=["/TTbar/Summer09-MC_31X_V3-v1/GEN-SIM-RAW"]
for dataset in datasets :
blocks=api.listBlocks(dataset)
for ablock in blocks:
data=api.listDatasetContents(dataset, ablock["Name"])
#print data
print "-- SQL Statements for Dataset : %s and Block : %s " % (dataset, ablock["Name"])
class Handler (xml.sax.handler.ContentHandler):
def __init__(self):
self.sqls={}
self.sqls['paths']=[]
self.sqls['storage_element']=[]
self.sqls['block_storage_elements']=[]
self.sqls['file']=[]
self.sqls['app_version']=[]
self.sqls['app_executable_name']=[]
self.sqls['ps_hash']=[]
from DBSAPI.dbsApi import DbsApi
from DBSAPI.dbsException import *
from DBSAPI.dbsApiException import *
from DBSAPI.dbsOptions import DbsOptionParser
try:
optManager = DbsOptionParser()
(opts, args) = optManager.getOpt()
#print opts.__dict__
api = DbsApi(opts.__dict__)
#for block in api.listBlocks("/TestPrimary_001_20070315_03h12m26s/TestProcessed_20070315_03h12m26s/GEN-SIM"):
#for block in api.listBlocks("/test_primary_001/TestProcessedDS001/GEN-SIM"):
#for block in api.listBlocks("/chi1/CMSSW_1_6_7-CSA07-3268/GEN-SIM-DIGI-RAW"):
#for block in api.listBlocks("/dataset_PD_110/CRUZET3-v1-unmerged/RAW"):
for block in api.listBlocks(
"/RelValSingleMuPt10/CMSSW_2_1_2_IDEAL_V6_v3/GEN-SIM-RECO"):
#for block in api.listBlocks(block_name="/test_primary_001*"):
#for block in api.listBlocks("", "/TestPrimary_001_20070315_02h26m11s/TestProcessed_20070315_02h26m11s/GEN-SIM#016712"):
#for block in api.listBlocks("/test_primary_001/TestProcessedDS001/GEN-SIM", "/test_primary_001/TestProcessedDS001/GEN*"):
#for block in api.listBlocks("/TestPrimary_001_20070315_02h53m32s/TestPrimary_001_20070315_02h53m32s/GEN-SIM"):
#print "%s %s" % (block['Name'], block['StorageElementList'])
#print " %s" % block['Name']
print " %s" % block
except DbsApiException, ex:
print "Caught API Exception %s: %s " % (ex.getClassName(),
ex.getErrorMessage())
if ex.getErrorCode() not in (None, ""):
print "DBS Exception Error Code: ", ex.getErrorCode()
class DBSWriter:
"""
_DBSWriter_
General API for writing data to DBS
"""
def __init__(self, url, **contact):
args = { "url" : url, "level" : 'ERROR'}
args.update(contact)
try:
self.dbs = DbsApi(args)
self.args = args
self.version = args.get('version', None)
self.globalDBSUrl = args.get('globalDBSUrl', None)
self.globalVersion = args.get('globalVersion', None)
if self.globalDBSUrl:
globalArgs = {'url': url, 'level': 'ERROR'}
globalArgs.update(contact)
self.globalDBS = DbsApi(globalArgs)
except DbsException as ex:
msg = "Error in DBSWriterError with DbsApi\n"
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
self.reader = DBSReader(**args)
def createDatasets(self, workflowSpec):
"""
_createDatasets_
Create All the output datasets found in the workflow spec instance
provided
"""
try:
workflowSpec.payload.operate(
_CreateDatasetOperator(self.dbs, workflowSpec)
)
except DbsException as ex:
msg = "Error in DBSWriter.createDatasets\n"
msg += "For Workflow: %s\n" % workflowSpec.workflowName()
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
return
def insertFilesForDBSBuffer(self, files, procDataset, algos,
jobType = "NotMerge", insertDetectorData = False,
maxFiles = 100, maxSize = 99999999, timeOut = None,
fileCommitLength = 5):
"""
_insertFiles_
list of files inserted in DBS
"""
#TODO: Whats the purpose of insertDetectorData
if len(files) < 1:
return
affectedBlocks = []
insertFiles = []
addedRuns=[]
pnn = None
#Get the algos in insertable form
# logging.error("About to input algos")
# logging.error(algos)
ialgos = [DBSWriterObjects.createAlgorithmForInsert(dict(algo)) for algo in algos ]
#print ialgos
for outFile in files:
# //
# // Convert each file into a DBS File object
#//
lumiList = []
#Somehing similar should be the real deal when multiple runs/lumi could be returned from wmbs file
for runlumiinfo in outFile.getRuns():
lrun=long(runlumiinfo.run)
run = DbsRun(
RunNumber = lrun,
NumberOfEvents = 0,
NumberOfLumiSections = 0,
TotalLuminosity = 0,
StoreNumber = 0,
StartOfRun = 0,
EndOfRun = 0,
)
#Only added if not added by another file in this loop, why waste a call to DBS
if lrun not in addedRuns:
self.dbs.insertRun(run)
addedRuns.append(lrun) #save it so we do not try to add it again to DBS
logging.debug("run %s added to DBS " % str(lrun))
for alsn in runlumiinfo:
lumi = DbsLumiSection(
LumiSectionNumber = long(alsn),
StartEventNumber = 0,
EndEventNumber = 0,
LumiStartTime = 0,
LumiEndTime = 0,
RunNumber = lrun,
)
lumiList.append(lumi)
logging.debug("lumi list created for the file")
dbsfile = DbsFile(
#Checksum = str(outFile['cksum']),
NumberOfEvents = outFile['events'],
LogicalFileName = outFile['lfn'],
FileSize = int(outFile['size']),
Status = "VALID",
ValidationStatus = 'VALID',
FileType = 'EDM',
Dataset = procDataset,
TierList = DBSWriterObjects.makeTierList(procDataset['Path'].split('/')[3]),
AlgoList = ialgos,
LumiList = lumiList,
ParentList = outFile.getParentLFNs(),
#BranchHash = outFile['BranchHash'],
)
#Set checksums by hand
#dbsfile['Checksum'] = 0 #Set a default?
for entry in outFile['checksums'].keys():
#This should be a dictionary with a cktype key and cksum value
if entry.lower() == 'cksum':
dbsfile['Checksum'] = str(outFile['checksums'][entry])
elif entry.lower() == 'adler32':
dbsfile['Adler32'] = str(outFile['checksums'][entry])
elif entry.lower() == 'md5':
dbsfile['Md5'] = str(outFile['checksums'][entry])
#This check comes from ProdAgent, not sure if its required
if len(outFile["locations"]) > 0:
pnn = list(outFile["locations"])[0]
logging.debug("PNN associated to file is: %s"%pnn)
else:
msg = "Error in DBSWriter.insertFiles\n"
msg += "No PNN associated to file"
#print "FAKING seName for now"
#seName="cmssrm.fnal.gov"
raise DBSWriterError(msg)
insertFiles.append(dbsfile)
# //Processing Jobs:
# // Insert the lists of sorted files into the appropriate
#// fileblocks
sumSize = 0
sumFiles = 0
tmpFiles = []
blockList = []
#First, get the block. See if the block already exists
try:
fileBlock = DBSWriterObjects.getDBSFileBlock(
self.dbs,
procDataset,
pnn)
fileBlock['files'] = []
#if not fileBlock in affectedBlocks:
# affectedBlocks.append(fileBlock)
except DbsException as ex:
msg = "Error in DBSWriter.insertFilesForDBSBuffer\n"
msg += "Cannot retrieve FileBlock for dataset:\n"
msg += " %s\n" % procDataset['Path']
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
filesToCommit = []
for file in insertFiles:
# First see if the block is full
if self.manageFileBlock(fileBlock = fileBlock, maxFiles = maxFiles,
maxSize = maxSize, timeOut = timeOut, algos = ialgos,
filesToCommit = filesToCommit, procDataset = procDataset):
fileBlock['OpenForWriting'] = 0
if not fileBlock in affectedBlocks:
affectedBlocks.append(fileBlock)
# Then we need a new block
try:
fileBlock = DBSWriterObjects.getDBSFileBlock(
self.dbs,
procDataset,
pnn)
fileBlock['files'] = []
except DbsException as ex:
msg = "Error in DBSWriter.insertFilesForDBSBuffer\n"
msg += "Cannot retrieve FileBlock for dataset:\n"
msg += " %s\n" % procDataset['Path']
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
fileBlock['files'].append(file['LogicalFileName'])
filesToCommit.append(file)
if len(filesToCommit) >= fileCommitLength:
# Only commit the files if there are more of them then the maximum length
try:
self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
filesToCommit = []
logging.debug("Inserted files: %s to FileBlock: %s" \
% ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
msg += " %s\n" % ([ x['LogicalFileName'] for x in insertFiles ],)
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
if len(filesToCommit) > 0:
try:
self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
filesToCommit = []
logging.debug("Inserted files: %s to FileBlock: %s" \
% ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
msg += " %s\n" % ([ x['LogicalFileName'] for x in insertFiles ],)
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
if not fileBlock in affectedBlocks:
affectedBlocks.append(fileBlock)
## Do bulk inserts now for DBS
#filesToCommit = []
#count = 0
#count2 = 0
#for file in insertFiles:
# count += 1
# #Try and close the box
# logging.error("Should have a file")
# logging.error(len(filesToCommit))
# count2 += len(filesToCommit)
# if self.manageFileBlock(fileBlock = fileBlock, maxFiles = maxFiles,
# maxSize = maxSize, timeOut = timeOut, algos = ialgos,
# filesToCommit = filesToCommit, procDataset = procDataset):
# fileBlock['OpenForWriting'] = '0'
# if not fileBlock in affectedBlocks:
# affectedBlocks.append(fileBlock)
#
#
#
# # Then we need a new block
# try:
# fileBlock = DBSWriterObjects.getDBSFileBlock(
# self.dbs,
# procDataset,
# seName)
# fileBlock['files'] = []
# except DbsException, ex:
# msg = "Error in DBSWriter.insertFilesForDBSBuffer\n"
# msg += "Cannot retrieve FileBlock for dataset:\n"
# msg += " %s\n" % procDataset['Path']
# msg += "%s\n" % formatEx(ex)
# raise DBSWriterError(msg)
# #At this point, we should commit the block as is
# fileBlock['files'].append(file['LogicalFileName'])
# if jobType == "MergeSpecial":
# for file in fileList:
# file['Block'] = fileBlock
# msg="calling: self.dbs.insertMergedFile(%s, %s)" % (str(file['ParentList']),str(file))
# logging.debug(msg)
# try:
# #
# #
# # NOTE To Anzar From Anzar (File cloning as in DBS API can be done here and then I can use Bulk insert on Merged files as well)
# self.dbs.insertMergedFile(file['ParentList'],
# file)
#
# except DbsException, ex:
# msg = "Error in DBSWriter.insertFiles\n"
# msg += "Cannot insert merged file:\n"
# msg += " %s\n" % file['LogicalFileName']
# msg += "%s\n" % formatEx(ex)
# raise DBSWriterError(msg)
# logging.debug("Inserted merged file: %s to FileBlock: %s"%(file['LogicalFileName'],fileBlock['Name']))
# else:
# filesToCommit.append(file)
# if len(filesToCommit) >= fileCommitLength:
# # Only commit the files if there are more of them then the maximum length
# try:
# logging.error("About to commit %i files" %(len(filesToCommit)))
# count2 += len(filesToCommit)
# self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
# filesToCommit = []
# logging.debug("Inserted files: %s to FileBlock: %s" \
# % ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
#
# except DbsException, ex:
# msg = "Error in DBSWriter.insertFiles\n"
# msg += "Cannot insert processed files:\n"
# msg += " %s\n" % ([ x['LogicalFileName'] for x in insertFiles ],)
# msg += "%s\n" % formatEx(ex)
# raise DBSWriterError(msg)
#
#
#
#
## If we still have files to commit, commit them
#logging.error("Got to the end of the loop")
#logging.error(len(filesToCommit))
#logging.error(count2)
#if len(filesToCommit) > 0:
# try:
# logging.error("About to insert some files")
# self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
# filesToCommit = []
# logging.debug("Inserted files: %s to FileBlock: %s" \
# % ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
#
# except DbsException, ex:
# msg = "Error in DBSWriter.insertFiles\n"
# msg += "Cannot insert processed files:\n"
# msg += " %s\n" % ([ x['LogicalFileName'] for x in insertFiles ],)
# msg += "%s\n" % formatEx(ex)
# raise DBSWriterError(msg)
if not fileBlock in affectedBlocks:
affectedBlocks.append(fileBlock)
return list(affectedBlocks)
def insertFiles(self, fwkJobRep, insertDetectorData = False):
"""
_insertFiles_
Process the files in the FwkJobReport instance and insert
them into the associated datasets
A list of affected fileblock names is returned both for merged
and unmerged fileblocks. Only merged blocks will have to be managed.
#for merged file
#blocks to facilitate management of those blocks.
#This list is not populated for processing jobs since we dont really
#care about the processing job blocks.
"""
insertLists = {}
orderedHashes = []
affectedBlocks = set()
if len(fwkJobRep.files)<=0:
msg = "Error in DBSWriter.insertFiles\n"
msg += "No files found in FrameWorkJobReport for:\n"
msg += "==> JobSpecId: %s"%fwkJobRep.jobSpecId
msg += " Workflow: %s"%fwkJobRep.workflowSpecId
raise DBSWriterError(msg)
for outFile in fwkJobRep.sortFiles():
# //
# // Convert each file into a DBS File object
#//
pnn = None
if "PNN" in outFile:
if outFile['PNN'] :
pnn = outFile['PNN']
logging.debug("PNN associated to file is: %s"%pnn)
## remove the fallback to site se-name if no SE is associated to File
## because it's likely that there is some stage out problem if there
## is no SEName associated to the file.
# if not seName:
# if fwkJobRep.siteDetails.has_key("se-name"):
# seName = fwkJobRep.siteDetails['se-name']
# seName = str(seName)
# logging.debug("site SEname: %s"%seName)
if not pnn:
msg = "Error in DBSWriter.insertFiles\n"
msg += "No PNN associated to files in FrameWorkJobReport for "
# msg += "No SEname found in FrameWorkJobReport for "
msg += "==> JobSpecId: %s"%fwkJobRep.jobSpecId
msg += " Workflow: %s"%fwkJobRep.workflowSpecId
raise DBSWriterError(msg)
try:
if ( insertDetectorData ):
dbsFiles = DBSWriterObjects.createDBSFiles(outFile,
fwkJobRep.jobType,
self.dbs)
else:
dbsFiles = DBSWriterObjects.createDBSFiles(outFile,
fwkJobRep.jobType)
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles:\n"
msg += "Error creating DbsFile instances for file:\n"
msg += "%s\n" % outFile['LFN']
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
if len(dbsFiles)<=0:
msg="No DbsFile instances created. Not enough info in the FrameWorkJobReport for"
msg += "==> JobSpecId: %s"%fwkJobRep.jobSpecId
msg += " Workflow: %s"%fwkJobRep.workflowSpecId
raise DBSWriterError(msg)
for f in dbsFiles:
datasetName = makeDBSDSName(f)
hashName = "%s-%s" % (pnn, datasetName)
if hashName not in insertLists:
insertLists[hashName] = _InsertFileList(pnn,
datasetName)
insertLists[hashName].append(f)
if not orderedHashes.count(hashName):
orderedHashes.append(hashName)
# //Processing Jobs:
# // Insert the lists of sorted files into the appropriate
#// fileblocks
for hash in orderedHashes:
fileList = insertLists[hash]
procDataset = fileList[0]['Dataset']
try:
fileBlock = DBSWriterObjects.getDBSFileBlock(
self.dbs,
procDataset,
fileList.pnn)
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot retrieve FileBlock for dataset:\n"
msg += " %s\n" % procDataset
# msg += "In Storage Element:\n %s\n" % fileList.seName
msg += "In PNN:\n %s\n" % fileList.pnn
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
if fwkJobRep.jobType == "Merge":
# //
# // Merge files
#//
for mergedFile in fileList:
mergedFile['Block'] = fileBlock
affectedBlocks.add(fileBlock['Name'])
msg="calling: self.dbs.insertMergedFile(%s, %s)" % (str(mergedFile['ParentList']),str(mergedFile))
logging.debug(msg)
try:
self.dbs.insertMergedFile(mergedFile['ParentList'],
mergedFile)
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert merged file:\n"
msg += " %s\n" % mergedFile['LogicalFileName']
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
logging.debug("Inserted merged file: %s to FileBlock: %s"%(mergedFile['LogicalFileName'],fileBlock['Name']))
else:
# //
# // Processing files
#//
affectedBlocks.add(fileBlock['Name'])
msg="calling: self.dbs.insertFiles(%s, %s, %s)" % (str(procDataset),str(list(fileList)),str(fileBlock))
logging.debug(msg)
try:
self.dbs.insertFiles(procDataset, list(fileList),
fileBlock)
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
msg += " %s\n" % (
[ x['LogicalFileName'] for x in fileList ],
)
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
logging.debug("Inserted files: %s to FileBlock: %s"%( ([ x['LogicalFileName'] for x in fileList ]),fileBlock['Name']))
return list(affectedBlocks)
def manageFileBlock(self, fileBlock, maxFiles = 100, maxSize = None,
timeOut = None, algos = [], filesToCommit = [],
procDataset = None):
"""
_manageFileBlock_
Check to see wether the fileblock with the provided name
is closeable based on number of files or total size.
If the block equals or exceeds wither the maxFiles or maxSize
parameters, close the block and return True, else do nothing and
return False
"""
# //
# // Check that the block exists, and is open before we close it
#//
fileblockName = fileBlock['Name']
blockInstance = self.dbs.listBlocks(block_name=fileblockName)
if len(blockInstance) > 1:
msg = "Multiple Blocks matching name: %s\n" % fileblockName
msg += "Unable to manage file block..."
raise DBSWriterError(msg)
if len(blockInstance) == 0:
msg = "Block name %s not found\n" % fileblockName
msg += "Cant manage a non-existent fileblock"
raise DBSWriterError(msg)
blockInstance = blockInstance[0]
isClosed = blockInstance.get('OpenForWriting', '1')
if isClosed != '1':
msg = "Block %s already closed" % fileblockName
logging.warning(msg)
# Now we need to commit files
if len(filesToCommit) > 0:
try:
self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
filesToCommit = []
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
raise DBSWriterError(msg)
# Attempting to migrate to global
if self.globalDBSUrl:
self.dbs.dbsMigrateBlock(srcURL = self.args['url'],
dstURL = self.globalDBSUrl,
block_name = fileblockName,
srcVersion = self.version,
dstVersion = self.globalVersion,
)
#for algo in algos:
# self.globalDBS.insertAlgoInPD(dataset = get_path(fileblockName.split('#')[0]),
# algorithm = algo)
logging.info("Migrated block %s to global due to pre-closed status" %(fileblockName))
else:
logging.error("Should've migrated block %s because it was already closed, but didn't" % (fileblockName))
return True
# //
# // We have an open block, sum number of files and file sizes
#//
#fileCount = int(blockInstance.get('NumberOfFiles', 0))
fileCount = len(fileBlock['files'])
totalSize = float(blockInstance.get('BlockSize', 0))
msg = "Fileblock: %s\n ==> Size: %s Files: %s\n" % (
fileblockName, totalSize, fileCount)
logging.warning(msg)
# //
# // Test close block conditions
#//
closeBlock = False
if timeOut:
if int(time.time()) - int(blockInstance['CreationDate']) > timeOut:
closeBlock = True
msg = "Closing Block based on timeOut: %s" % fileblockName
logging.debug(msg)
if fileCount >= maxFiles:
closeBlock = True
msg = "Closing Block Based on files: %s" % fileblockName
logging.debug(msg)
if maxSize != None:
if totalSize >= maxSize:
closeBlock = True
msg = "Closing Block Based on size: %s" % fileblockName
logging.debug(msg)
if closeBlock:
# Now we need to commit files
if len(filesToCommit) > 0:
try:
self.dbs.insertFiles(procDataset, filesToCommit, fileBlock)
filesToCommit = []
#logging.debug("Inserted files: %s to FileBlock: %s" \
# % ( ([ x['LogicalFileName'] for x in insertFiles ]),fileBlock['Name']))
except DbsException as ex:
msg = "Error in DBSWriter.insertFiles\n"
msg += "Cannot insert processed files:\n"
#msg += " %s\n" % ([ x['LogicalFileName'] for x in insertFiles ],)
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
# //
# // Close the block
#//
self.dbs.closeBlock(
DBSWriterObjects.createDBSFileBlock(fileblockName)
)
if self.globalDBSUrl:
self.dbs.dbsMigrateBlock(srcURL = self.args['url'],
dstURL = self.globalDBSUrl,
block_name = fileblockName,
srcVersion = self.version,
dstVersion = self.globalVersion
)
for algo in algos:
pass
#self.globalDBS.insertAlgoInPD(dataset = get_path(fileblockName.split('#')[0]),
# algorithm = algo)
logging.info("Migrated block %s to global" %(fileblockName))
else:
logging.error("Should've migrated block %s, but didn't" % (fileblockName))
return closeBlock
def migrateDatasetBlocks(self, inputDBSUrl, datasetPath, blocks):
"""
_migrateDatasetBlocks_
Migrate the list of fileblocks provided by blocks, belonging
to the dataset specified by the dataset path to this DBS instance
from the inputDBSUrl provided
- *inputDBSUrl* : URL for connection to input DBS
- *datasetPath* : Name of dataset in input DBS (must exist in input
DBS)
- *blocks* : list of block names to be migrated (must exist
in input DBS)
"""
if len(blocks) == 0:
msg = "FileBlocks not provided.\n"
msg += "You must provide the name of at least one fileblock\n"
msg += "to be migrated"
raise DBSWriterError(msg)
# //
# // Hook onto input DBSUrl and verify that the dataset & blocks
#// exist
reader = DBSReader(inputDBSUrl)
inputBlocks = reader.listFileBlocks(datasetPath)
for block in blocks:
# //
# // Test block exists at source
#//
if block not in inputBlocks:
msg = "Block name:\n ==> %s\n" % block
msg += "Not found in input dataset:\n ==> %s\n" % datasetPath
msg += "In DBS Instance:\n ==> %s\n" % inputDBSUrl
raise DBSWriterError(msg)
# //
# // Test block does not exist in target
#//
if self.reader.blockExists(block):
# //
# // block exists
#// If block is closed dont attempt transfer
if not self.reader.blockIsOpen(block):
msg = "Block already exists in target DBS and is closed:\n"
msg += " ==> %s\n" % block
msg += "Skipping Migration of that block"
logging.warning(msg)
continue
try:
xferData = reader.dbs.listDatasetContents(datasetPath, block)
except DbsException as ex:
msg = "Error in DBSWriter.migrateDatasetBlocks\n"
msg += "Could not read content of dataset:\n ==> %s\n" % (
datasetPath,)
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
xferData = _remapBlockParentage(datasetPath, xferData)
try:
self.dbs.insertDatasetContents(xferData)
except DbsException as ex:
msg = "Error in DBSWriter.migrateDatasetBlocks\n"
msg += "Could not write content of dataset:\n ==> %s\n" % (
datasetPath,)
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
del xferData
return
def importDatasetWithExistingParents(self, sourceDBS, sourceDatasetPath, targetDBS,
onlyClosed = True):
"""
_importDataset_
Import a dataset into the local scope DBS.
It complains if the parent dataset ar not there!!
- *sourceDBS* : URL for input DBS instance
- *sourceDatasetPath* : Dataset Path to be imported
- *targetDBS* : URL for DBS to have dataset imported to
"""
reader = DBSReader(sourceDBS)
inputBlocks = reader.getFileBlocksInfo(sourceDatasetPath, onlyClosed, locations = False)
for inputBlock in inputBlocks:
block = inputBlock['Name']
# //
# // Test block does not exist in target
#//
if self.reader.blockExists(block):
# //
# // block exists
#// If block is closed dont attempt transfer
if not str(inputBlock['OpenForWriting']) != '1':
msg = "Block already exists in target DBS and is closed:\n"
msg += " ==> %s\n" % block
msg += "Skipping Import of that block"
logging.warning(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDatasetWithExistingParents\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
logging.info("Update block locations to:")
for pnn in locations:
self.dbs.addReplicaToBlock(block,pnn)
logging.info(pnn)
continue
try:
xferData = reader.dbs.listDatasetContents(
sourceDatasetPath, block
)
except DbsException as ex:
msg = "Error in DBSWriter.importDatasetWithExistingParents\n"
msg += "Could not read content of dataset:\n ==> %s\n" % (
sourceDatasetPath,)
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
try:
self.dbs.insertDatasetContents(xferData)
except DbsException as ex:
msg = "Error in DBSWriter.importDatasetWithExistingParents\n"
msg += "Could not write content of dataset:\n ==> %s\n" % (
sourceDatasetPath,)
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
del xferData
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDatasetWithExistingParents\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
for pnn in locations:
self.dbs.addReplicaToBlock(block,pnn)
return
def importDataset(self, sourceDBS, sourceDatasetPath, targetDBS,
onlyClosed = True):
"""
_importDataset_
Import a dataset into the local scope DBS with full parentage hirerarchy
(at least not slow because branches info is dropped)
- *sourceDBS* : URL for input DBS instance
- *sourceDatasetPath* : Dataset Path to be imported
- *targetDBS* : URL for DBS to have dataset imported to
"""
reader = DBSReader(sourceDBS)
inputBlocks = reader.getFileBlocksInfo(sourceDatasetPath, onlyClosed, locations = False)
blkCounter=0
for inputBlock in inputBlocks:
block = inputBlock['Name']
# //
# // Test block does not exist in target
#//
blkCounter=blkCounter+1
msg="Importing block %s of %s: %s " % (blkCounter,len(inputBlocks),block)
logging.debug(msg)
if self.reader.blockExists(block):
# //
# // block exists
#// If block is closed dont attempt transfer
if str(inputBlock['OpenForWriting']) != '1':
msg = "Block already exists in target DBS and is closed:\n"
msg += " ==> %s\n" % block
msg += "Skipping Import of that block"
logging.warning(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDataset\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
logging.info("Update block locations to:")
for pnn in locations:
self.dbs.addReplicaToBlock(block,pnn)
logging.info(pnn)
continue
try:
self.dbs.migrateDatasetContents(sourceDBS, targetDBS, sourceDatasetPath, block_name=block, noParentsReadOnly = False)
except DbsException as ex:
msg = "Error in DBSWriter.importDataset\n"
msg += "Could not write content of dataset:\n ==> %s\n" % (
sourceDatasetPath,)
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDataset\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
for pnn in locations:
self.dbs.addReplicaToBlock(block,pnn)
return
def importDatasetWithoutParentage(self, sourceDBS, sourceDatasetPath, targetDBS,
onlyClosed = True):
"""
_importDataset_
Import a dataset into the local scope DBS with one level parentage,
however it has severe limitation on its use due to the "ReadOnly" concept.
- *sourceDBS* : URL for input DBS instance
- *sourceDatasetPath* : Dataset Path to be imported
- *targetDBS* : URL for DBS to have dataset imported to
"""
reader = DBSReader(sourceDBS)
inputBlocks = reader.getFileBlocksInfo(sourceDatasetPath, onlyClosed, locations = False)
for inputBlock in inputBlocks:
block = inputBlock['Name']
# //
# // Test block does not exist in target
#//
if self.reader.blockExists(block):
# //
# // block exists
#// If block is closed dont attempt transfer
if str(inputBlock['OpenForWriting']) != '1':
msg = "Block already exists in target DBS and is closed:\n"
msg += " ==> %s\n" % block
msg += "Skipping Import of that block"
logging.warning(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDatasetWithoutParentage\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
logging.info("Update block locations to:")
for pnn in locations:
self.dbs.addReplicaToBlock(block,pnn)
logging.info(pnn)
continue
try:
self.dbs.migrateDatasetContents(sourceDBS, targetDBS, sourceDatasetPath, block_name=block, noParentsReadOnly = True )
except DbsException as ex:
msg = "Error in DBSWriter.importDatasetWithoutParentage\n"
msg += "Could not write content of dataset:\n ==> %s\n" % (
sourceDatasetPath,)
msg += "Block name:\n ==> %s\n" % block
msg += "%s\n" % formatEx(ex)
raise DBSWriterError(msg)
locations = reader.listFileBlockLocation(block)
# only empty file blocks can have no location
if not locations and str(inputBlock['NumberOfFiles']) != "0":
msg = "Error in DBSWriter.importDatasetWithoutParentage\n"
msg += "Block has no locations defined: %s" % block
raise DBSWriterError(msg)
for pnn in locations:
self.dbs.addReplicaToBlock(block,pnn)
return
def getOutputDatasetsWithPSet(payloadNode):
"""
_getOutputDatasetsWithPSet_
Extract all the information about output datasets from the
payloadNode object provided, including the {{}} format PSet cfg
Returns a list of DatasetInfo objects including App details
from the node.
"""
result = []
for item in payloadNode._OutputDatasets:
resultEntry = DatasetInfo()
resultEntry.update(item)
resultEntry["ApplicationName"] = payloadNode.application['Executable']
resultEntry["ApplicationProject"] = payloadNode.application['Project']
resultEntry["ApplicationVersion"] = payloadNode.application['Version']
resultEntry["ApplicationFamily"] = item.get("OutputModuleName", "AppFamily")
try:
config = payloadNode.cfgInterface
psetStr = config.originalContent()
resultEntry['PSetContent'] = psetStr
except Exception as ex:
resultEntry['PSetContent'] = None
result.append(resultEntry)
return _sortDatasets(result)
