def loadFiles(self):
"""
_loadFiles_
Load all files that need to be loaded.
I will do this by DAS for now to break
the monstrous calls down into smaller chunks.
"""
# Grab all the Dataset-Algo combindations
dasList = self.dbsUtil.findUploadableDAS()
if len(dasList) < 1:
# Then there's nothing to do
return []
readyBlocks = []
for dasInfo in dasList:
dasID = dasInfo['DAS_ID']
# Get the files
try:
loadedFiles = self.dbsUtil.findUploadableFilesByDAS(das = dasID)
except WMException:
raise
except Exception, ex:
msg = "Unhandled exception while loading uploadable files for DAS.\n"
msg += str(ex)
logging.error(msg)
logging.debug("DAS being loaded: %s\n" % dasID)
raise DBSUploadException(msg)
# Get the blocks
if not dasID in self.dasCache.keys():
# Then we have a new DAS
# Add it
self.dasCache[dasID] = {}
dasBlocks = self.dasCache.get(dasID)
# Sort the files and blocks by location
fileDict = sortListByKey(input = loadedFiles, key = 'locations')
# Now we have both files and blocks
# We need a sorting algorithm of sorts...
# Now add each file
for location in fileDict.keys():
files = fileDict.get(location)
if len(files) < 1:
# Nothing to do here
continue
dasBlocks = self.dasCache[dasID].get(location, [])
if len(dasBlocks) > 0:
# Load from cache
currentBlock = self.blockCache.get(dasBlocks[0])
else:
blockname = '%s#%s' % (files[0]['datasetPath'], makeUUID())
currentBlock = DBSBlock(name = blockname,
location = location, das = dasID)
# Add the era info
currentBlock.setAcquisitionEra(era = dasInfo['AcquisitionEra'])
currentBlock.setProcessingVer(era = dasInfo['ProcessingVer'])
self.addNewBlock(block = currentBlock)
dasBlocks.append(currentBlock.getName())
for newFile in files:
if not newFile.get('block', 1) == None:
# Then this file already has a block
# It should be accounted for somewhere
# Or loaded with the block
continue
# Check if we can put files in this block
if not self.isBlockOpen(newFile = newFile,
block = currentBlock):
# Then we have to close the block and get a new one
currentBlock.status = 'Pending'
readyBlocks.append(currentBlock)
dasBlocks.remove(currentBlock.getName())
currentBlock = self.getBlock(newFile = newFile,
dasBlocks = dasBlocks,
location = location,
das = dasID)
currentBlock.setAcquisitionEra(era = dasInfo['AcquisitionEra'])
currentBlock.setProcessingVer(era = dasInfo['ProcessingVer'])
# Now deal with the file
currentBlock.addFile(dbsFile = newFile)
self.filesToUpdate.append({'filelfn': newFile['lfn'],
'block': currentBlock.getName()})
# Done with the location
readyBlocks.append(currentBlock)
def testDualUpload(self):
"""
_testDualUpload_
Verify that the dual upload mode works correctly.
"""
self.dbsApi = DbsApi(url = self.dbsUrl)
config = self.getConfig(dbs3UploadOnly = True)
dbsUploader = DBSUploadPoller(config = config)
dbsUtil = DBSBufferUtil()
# First test verifies that uploader will poll and then not do anything
# as the database is empty.
dbsUploader.algorithm()
acqEra = "Summer%s" % (int(time.time()))
parentFiles = self.createParentFiles(acqEra)
(moreParentFiles, childFiles) = \
self.createFilesWithChildren(parentFiles, acqEra)
allFiles = parentFiles + moreParentFiles
allBlocks = []
for i in range(4):
blockName = parentFiles[0]["datasetPath"] + "#" + makeUUID()
dbsBlock = DBSBlock(blockName, "malpaquet", 1)
dbsBlock.status = "Open"
dbsUtil.createBlocks([dbsBlock])
for file in allFiles[i * 5 : (i * 5) + 5]:
dbsBlock.addFile(file)
dbsUtil.setBlockFiles({"block": blockName, "filelfn": file["lfn"]})
if i < 2:
dbsBlock.status = "InDBS"
dbsUtil.updateBlocks([dbsBlock])
dbsUtil.updateFileStatus([dbsBlock], "InDBS")
allBlocks.append(dbsBlock)
blockName = childFiles[0]["datasetPath"] + "#" + makeUUID()
dbsBlock = DBSBlock(blockName, "malpaquet", 1)
dbsBlock.status = "InDBS"
dbsUtil.createBlocks([dbsBlock])
for file in childFiles:
dbsBlock.addFile(file)
dbsUtil.setBlockFiles({"block": blockName, "filelfn": file["lfn"]})
dbsUtil.updateFileStatus([dbsBlock], "InDBS")
dbsUploader.algorithm()
time.sleep(5)
dbsUploader.algorithm()
time.sleep(5)
self.verifyData(parentFiles[0]["datasetPath"], parentFiles)
# Change the status of the rest of the parent blocks so we can upload
# them and the children.
for dbsBlock in allBlocks:
dbsBlock.status = "InDBS"
dbsUtil.updateBlocks([dbsBlock])
dbsUploader.algorithm()
time.sleep(5)
self.verifyData(parentFiles[0]["datasetPath"], parentFiles + moreParentFiles)
# Run the uploader one more time to upload the children.
dbsUploader.algorithm()
time.sleep(5)
self.verifyData(childFiles[0]["datasetPath"], childFiles)
return
block.FillFromDBSBuffer(blockInfo)
blockname = block.getName()
# Now we have to load files...
try:
files = self.dbsUtil.loadFilesByBlock(blockname = blockname)
except WMException:
raise
except Exception, ex:
msg = "Unhandled exception while loading files for existing blocks.\n"
msg += str(ex)
logging.error(msg)
logging.debug("Blocks being loaded: %s\n" % blockname)
raise DBSUploadException(msg)
for file in files:
block.addFile(file)
# Add to the cache
self.addNewBlock(block = block)
# All blocks should now be loaded and present
# in both the block cache (which has all the info)
# and the dasCache (which is a list of name pointers
# to the keys in the block cache).
return
def loadFiles(self):
"""
try:
files = self.dbsUtil.loadFilesByBlock(blockname=blockname)
logging.info("Have %i files for block %s" %
(len(files), blockname))
except WMException:
raise
except Exception, ex:
msg = "Unhandled exception while loading files for existing blocks.\n"
msg += str(ex)
logging.error(msg)
logging.debug("Blocks being loaded: %s\n" % blockname)
raise DBSUploadException(msg)
# Add the loaded files to the block
for file in files:
block.addFile(file, self.datasetType, self.primaryDatasetType)
# Add to the cache
self.addNewBlock(block=block)
# All blocks should now be loaded and present
# in both the block cache (which has all the info)
# and the dasCache (which is a list of name pointers
# to the keys in the block cache).
return
def loadFiles(self):
"""
_loadFiles_
Load all files that need to be loaded. I will do this by DAS for now to
def testDualUpload(self):
"""
_testDualUpload_
Verify that the dual upload mode works correctly.
"""
self.dbsApi = DbsApi(url=self.dbsUrl)
dbsUploader = DBSUploadPoller(config=config)
dbsUtil = DBSBufferUtil()
# First test verifies that uploader will poll and then not do anything
# as the database is empty.
dbsUploader.algorithm()
acqEra = "Summer%s" % (int(time.time()))
parentFiles = self.createParentFiles(acqEra)
(moreParentFiles, childFiles) = \
self.createFilesWithChildren(parentFiles, acqEra)
allFiles = parentFiles + moreParentFiles
allBlocks = []
for i in range(4):
DBSBufferDataset(parentFiles[0]["datasetPath"]).create()
blockName = parentFiles[0]["datasetPath"] + "#" + makeUUID()
dbsBlock = DBSBlock(blockName,
location="malpaquet",
das=None,
workflow=None)
dbsBlock.status = "Open"
dbsBlock.setDataset(parentFiles[0]["datasetPath"], 'data', 'VALID')
dbsUtil.createBlocks([dbsBlock])
for file in allFiles[i * 5:(i * 5) + 5]:
dbsBlock.addFile(file, 'data', 'VALID')
dbsUtil.setBlockFiles({
"block": blockName,
"filelfn": file["lfn"]
})
if i < 2:
dbsBlock.status = "InDBS"
dbsUtil.updateBlocks([dbsBlock])
dbsUtil.updateFileStatus([dbsBlock], "InDBS")
allBlocks.append(dbsBlock)
DBSBufferDataset(childFiles[0]["datasetPath"]).create()
blockName = childFiles[0]["datasetPath"] + "#" + makeUUID()
dbsBlock = DBSBlock(blockName,
location="malpaquet",
das=None,
workflow=None)
dbsBlock.status = "InDBS"
dbsBlock.setDataset(childFiles[0]["datasetPath"], 'data', 'VALID')
dbsUtil.createBlocks([dbsBlock])
for file in childFiles:
dbsBlock.addFile(file, 'data', 'VALID')
dbsUtil.setBlockFiles({"block": blockName, "filelfn": file["lfn"]})
dbsUtil.updateFileStatus([dbsBlock], "InDBS")
dbsUploader.algorithm()
time.sleep(5)
dbsUploader.algorithm()
time.sleep(5)
self.verifyData(parentFiles[0]["datasetPath"], parentFiles)
# Change the status of the rest of the parent blocks so we can upload
# them and the children.
for dbsBlock in allBlocks:
dbsBlock.status = "InDBS"
dbsUtil.updateBlocks([dbsBlock])
dbsUploader.algorithm()
time.sleep(5)
self.verifyData(parentFiles[0]["datasetPath"],
parentFiles + moreParentFiles)
# Run the uploader one more time to upload the children.
dbsUploader.algorithm()
time.sleep(5)
self.verifyData(childFiles[0]["datasetPath"], childFiles)
return
# Now we have to load files...
try:
files = self.dbsUtil.loadFilesByBlock(blockname = blockname)
logging.info("Have %i files for block %s" % (len(files), blockname))
except WMException:
raise
except Exception, ex:
msg = "Unhandled exception while loading files for existing blocks.\n"
msg += str(ex)
logging.error(msg)
logging.debug("Blocks being loaded: %s\n" % blockname)
raise DBSUploadException(msg)
# Add the loaded files to the block
for file in files:
block.addFile(file, self.datasetType, self.primaryDatasetType)
# Add to the cache
self.addNewBlock(block = block)
# All blocks should now be loaded and present
# in both the block cache (which has all the info)
# and the dasCache (which is a list of name pointers
# to the keys in the block cache).
return
def loadFiles(self):
"""
_loadFiles_
def loadFiles(self):
"""
_loadFiles_
Load all files that need to be loaded. I will do this by DAS for now to
break the monstrous calls down into smaller chunks.
"""
# Grab all the Dataset-Algo combindations
dasList = self.dbsUtil.findUploadableDAS()
if len(dasList) < 1:
# Then there's nothing to do
return []
readyBlocks = []
for dasInfo in dasList:
dasID = dasInfo['DAS_ID']
# Get the files
try:
loadedFiles = self.dbsUtil.findUploadableFilesByDAS(das=dasID)
except WMException:
raise
except Exception, ex:
msg = "Unhandled exception while loading uploadable files for DAS.\n"
msg += str(ex)
logging.error(msg)
logging.debug("DAS being loaded: %s\n" % dasID)
raise DBSUploadException(msg)
# Get the blocks
if not dasID in self.dasCache.keys():
# Then we have a new DAS
# Add it
self.dasCache[dasID] = {}
dasBlocks = self.dasCache.get(dasID)
# Sort the files and blocks by location
fileDict = sortListByKey(input=loadedFiles, key='locations')
# Now add each file
for location in fileDict.keys():
files = fileDict.get(location)
if len(files) < 1:
# Nothing to do here
continue
dasBlocks = self.dasCache[dasID].get(location, [])
if len(dasBlocks) > 0:
# Load from cache
currentBlock = self.blockCache.get(dasBlocks[0])
else:
blockname = '%s#%s' % (files[0]['datasetPath'], makeUUID())
currentBlock = DBSBlock(name=blockname,
location=location,
das=dasID)
# Add the era info
currentBlock.setAcquisitionEra(
era=dasInfo['AcquisitionEra'])
currentBlock.setProcessingVer(
procVer=dasInfo['ProcessingVer'])
self.addNewBlock(block=currentBlock)
dasBlocks.append(currentBlock.getName())
for newFile in files:
if not newFile.get('block', 1) == None:
# Then this file already has a block
# It should be accounted for somewhere
# Or loaded with the block
continue
# Check if we can put files in this block
if not self.isBlockOpen(newFile=newFile,
block=currentBlock):
# Then we have to close the block and get a new one
currentBlock.status = 'Pending'
readyBlocks.append(currentBlock)
dasBlocks.remove(currentBlock.getName())
currentBlock = self.getBlock(newFile=newFile,
dasBlocks=dasBlocks,
location=location,
das=dasID)
currentBlock.setAcquisitionEra(
era=dasInfo['AcquisitionEra'])
currentBlock.setProcessingVer(
procVer=dasInfo['ProcessingVer'])
# Now deal with the file
currentBlock.addFile(dbsFile=newFile)
self.filesToUpdate.append({
'filelfn': newFile['lfn'],
'block': currentBlock.getName()
})
# Done with the location
readyBlocks.append(currentBlock)
try:
files = self.dbsUtil.loadFilesByBlock(blockname=blockname)
logging.info("Have %i files for block %s" %
(len(files), blockname))
except WMException:
raise
except Exception, ex:
msg = "Unhandled exception while loading files for existing blocks.\n"
msg += str(ex)
logging.error(msg)
logging.debug("Blocks being loaded: %s\n" % blockname)
raise DBSUploadException(msg)
# Add the loaded files to the block
for file in files:
block.addFile(file)
# Add to the cache
self.addNewBlock(block=block)
# All blocks should now be loaded and present
# in both the block cache (which has all the info)
# and the dasCache (which is a list of name pointers
# to the keys in the block cache).
return
def loadFiles(self):
"""
_loadFiles_
Load all files that need to be loaded. I will do this by DAS for now to
def loadBlocks(self):
"""
_loadBlocks_
Find all blocks; make sure they're in the cache
"""
openBlocks = self.dbsUtil.findOpenBlocks()
logging.info("These are the openblocks: %s" % openBlocks)
# Load them if we don't have them
blocksToLoad = []
for block in openBlocks:
if not block['blockname'] in self.blockCache.keys():
blocksToLoad.append(block['blockname'])
# Now load the blocks
try:
loadedBlocks = self.dbsUtil.loadBlocks(blocksToLoad)
logging.info("Loaded blocks: %s" % loadedBlocks)
except WMException:
raise
except Exception as ex:
msg = "Unhandled exception while loading blocks.\n"
msg += str(ex)
logging.error(msg)
logging.debug("Blocks to load: %s\n" % blocksToLoad)
raise DBSUploadException(msg)
for blockInfo in loadedBlocks:
das = blockInfo['DatasetAlgo']
loc = blockInfo['origin_site_name']
workflow = blockInfo['workflow']
block = DBSBlock(name = blockInfo['block_name'],
location = loc, das = das, workflow = workflow)
block.FillFromDBSBuffer(blockInfo)
blockname = block.getName()
# Now we have to load files...
try:
files = self.dbsUtil.loadFilesByBlock(blockname = blockname)
logging.info("Have %i files for block %s" % (len(files), blockname))
except WMException:
raise
except Exception as ex:
msg = "Unhandled exception while loading files for existing blocks.\n"
msg += str(ex)
logging.error(msg)
logging.debug("Blocks being loaded: %s\n" % blockname)
raise DBSUploadException(msg)
# Add the loaded files to the block
for file in files:
block.addFile(file, self.datasetType, self.primaryDatasetType)
# Add to the cache
self.addNewBlock(block = block)
# All blocks should now be loaded and present
# in both the block cache (which has all the info)
# and the dasCache (which is a list of name pointers
# to the keys in the block cache).
return
def loadBlocks(self):
"""
_loadBlocks_
Find all blocks; make sure they're in the cache
"""
openBlocks = self.dbsUtil.findOpenBlocks(self.dbs3UploadOnly)
logging.info("These are the openblocks: %s" % openBlocks)
# Load them if we don't have them
blocksToLoad = []
for block in openBlocks:
if not block['blockname'] in self.blockCache.keys():
blocksToLoad.append(block['blockname'])
# Now load the blocks
try:
loadedBlocks = self.dbsUtil.loadBlocks(blocksToLoad, self.dbs3UploadOnly)
logging.info("Loaded blocks: %s" % loadedBlocks)
except WMException:
raise
except Exception as ex:
msg = "Unhandled exception while loading blocks.\n"
msg += str(ex)
logging.error(msg)
logging.debug("Blocks to load: %s\n" % blocksToLoad)
raise DBSUploadException(msg)
for blockInfo in loadedBlocks:
das = blockInfo['DatasetAlgo']
loc = blockInfo['origin_site_name']
workflow = blockInfo['workflow']
block = DBSBlock(name = blockInfo['block_name'],
location = loc, das = das, workflow = workflow)
block.FillFromDBSBuffer(blockInfo)
blockname = block.getName()
# Now we have to load files...
try:
files = self.dbsUtil.loadFilesByBlock(blockname = blockname)
logging.info("Have %i files for block %s" % (len(files), blockname))
except WMException:
raise
except Exception as ex:
msg = "Unhandled exception while loading files for existing blocks.\n"
msg += str(ex)
logging.error(msg)
logging.debug("Blocks being loaded: %s\n" % blockname)
raise DBSUploadException(msg)
# Add the loaded files to the block
for file in files:
block.addFile(file, self.datasetType, self.primaryDatasetType)
# Add to the cache
self.addNewBlock(block = block)
# All blocks should now be loaded and present
# in both the block cache (which has all the info)
# and the dasCache (which is a list of name pointers
# to the keys in the block cache).
return
