class RunSync(object):
"""
Synchronize the replica of a given run at WIPAC-ORIG
the corresponding Rucio site.
"""
def __init__(self,
run,
originrse=DEFAULT_ORIGIN_RSE,
destrse=None,
scope=DEFAULT_SCOPE,
check=True,
lifetime=None,
dry_run=False,
container=None):
"""
:param dataset: Name of the PhEDEx dataset to synchronize with Rucio.
:param pnn: PhEDEx node name to filter on for replica information.
"""
self.run = run
self.originrse = originrse
self.destrse = destrse
self.scope = scope
self.check = check
self.lifetime = lifetime
self.dry_run = dry_run
self.container = container
self.rucio_datasets = {}
self.run_files = {}
self.existent_replica_files = {}
self.url = ''
self.gfal = Gfal2Context()
self.run_Number = None
self.get_run_Number()
self.files_storage = {}
self.get_global_url()
self.didc = DIDClient()
self.repc = ReplicaClient()
self.rulesClient = RuleClient()
# Right now obtaining the Metadata from the storage at WIPAC
# Hopefully in the future from JADE # TODO
self.get_run_Files()
self.get_rucio_metadata()
self.update_run_Files()
self.get_files_metadata()
def update_run_Files(self):
"""
Updating the run files wiht only the files that have not been registered
"""
for f in self.existent_replica_files:
file_name = f.split('/')[-1:][0]
if file_name in self.run_files:
print("File: %s already registered. Skipping it" % file_name)
self.run_files.pop(file_name)
def get_files_metadata(self):
for f in self.run_files:
if self.run + '/' + f not in self.existent_replica_files:
self.obtain_metadata(f)
print("Metadat initialization done")
def obtain_metadata(self, filename):
"""
Get the size and checksum for every file in the run from the gftp server
"""
url = self.get_file_url(filename)
print("checking metadata for url %s" % url)
try:
size = self.gfal.stat(str(url)).st_size
adler32 = self.gfal.checksum(str(url), 'adler32')
print(
"got size and adler 32checksum of file: pfn=%s size=%s checksum=%s"
% (url, size, adler32))
self.run_files[filename] = {
'size': size,
'adler32': adler32,
'name': self.run + '/' + filename
}
except GError:
print("no file found at %s" % url)
return False
def get_file_url(self, filename):
return self.url + '/' + self.run + '/' + filename
def get_global_url(self):
"""
Return the base path of the rucio url
"""
print("Getting parameters for rse %s" % self.originrse)
rse = rsemgr.get_rse_info(self.originrse)
proto = rse['protocols'][0]
schema = proto['scheme']
prefix = proto['prefix'] + self.scope.replace('.', '/')
if schema == 'srm':
prefix = proto['extended_attributes']['web_service_path'] + prefix
url = schema + '://' + proto['hostname']
if proto['port'] != 0:
url = url + ':' + str(proto['port'])
self.url = url + prefix
print("Determined base url %s" % self.url)
def get_run_Number(self):
"""
Obtain the run number out of whole run IceCube/2016/filtered/level2pass2/0101/Run00127347
"""
print("Obtaining run number out of run(dataset): %s" % self.run)
self.run_Number = self.run.split("/")[-1]
print("Run number (dataset): %s" % self.run_Number)
def get_run_Files(self):
"""
Gets the list of files for a given run and their checksums from the storage
"""
self.run_url = self.url + '/' + self.run
print("Listin files from url : %s" % self.run_url)
run_files = []
try:
run_files = self.gfal.listdir(str(self.run_url))
except GError:
print("No files found at %s" % str(self.run_url))
print("Files found in storage:")
count = 0
for f in run_files:
if len(f) > 3:
if count < 5000:
self.run_files[f] = {}
count = count + 1
else:
break
def get_rucio_metadata(self):
"""
Gets the list of datasets at the Rucio RSE, the files, and the metadata.
"""
print(
"Initializing Rucio... getting the list of blocks and files at %s"
% self.originrse)
registered_datasets = self.repc.list_datasets_per_rse(self.originrse)
for dataset in registered_datasets:
self.rucio_datasets[dataset] = {}
replica_info = self.repc.list_replicas([{
"scope": self.scope,
"name": '/' + self.run_Number
}],
rse_expression="rse=%s" %
self.originrse)
replica_files = set()
for file_info in replica_info:
name = file_info['name']
if self.originrse in file_info['rses']:
replica_files.add(name)
self.existent_replica_files = replica_files
print("Rucio initialization done.")
def register(self):
"""
Create the container, the datasets and attach them to the container.
"""
print("Registering...")
self.register_dataset(self.run_Number)
self.register_replicas(self.run_files)
self.register_container(self.container)
self.attach_dataset_to_container(self.run_Number, self.container)
self.add_replica_rule(dataset=self.run_Number, destRSE=self.destrse)
def register_container(self, container):
"""
Registering the container
"""
print("Registering the container %s with scope: %s" %
(container, self.scope))
if container is None:
print('No container added, not registering any container')
return
if self.dry_run:
print('Dry run only, not registering the container')
return
try:
self.didc.add_container(scope=self.scope,
name=container,
lifetime=self.lifetime)
except DataIdentifierAlreadyExists:
print("Container %s already exists" % container)
except InvalidObject:
print("Problem with container name: %s" % container)
def attach_dataset_to_container(self, dataset, container):
"""
Attaching the dataset to a container
"""
print("Attaching dataset %s, to container: %s" % (dataset, container))
if container is None:
print('No container added, not registering dataset in container')
return
if self.dry_run:
print('Dry run only, not attaching dataset container')
return
try:
self.didc.attach_dids(scope=self.scope,
name=container,
dids=[{
'scope': self.scope,
'name': '/' + dataset
}])
except RucioException:
print("dataset already attached to container")
return
def register_dataset(self, run):
"""
Registering a dataset in the rucio database
"""
print("registering dataset %s" % run)
if self.dry_run:
print(' Dry run only. Not creating dataset.')
return
try:
self.didc.add_dataset(scope=self.scope,
name=run,
lifetime=self.lifetime)
except DataIdentifierAlreadyExists:
print(" Dataset %s already exists" % run)
def register_replicas(self, replicas):
"""
Register file replica.
"""
if not replicas:
return
print("registering files in Rucio: %s" %
", ".join([replicas[filemd]['name'] for filemd in replicas]))
if self.dry_run:
print(' Dry run only. Not registering files.')
return
try:
self.repc.add_replicas(rse=self.originrse,
files=[{
'scope':
self.scope,
'name':
replicas[filemd]['name'],
'adler32':
replicas[filemd]['adler32'],
'bytes':
replicas[filemd]['size'],
} for filemd in replicas])
print("Adding files to dataset: %s" % self.run_Number)
except InvalidObject:
print("Problem with file name does not match pattern")
for filemd in replicas:
try:
self.didc.attach_dids(scope=self.scope,
name=self.run_Number,
dids=[{
'scope': self.scope,
'name': replicas[filemd]['name']
}])
except FileAlreadyExists:
print("File already attached")
def add_replica_rule(self, destRSE, dataset):
"""
Create a replication rule for one dataset "Run" at an RSE
"""
print("Creating replica rule for dataset %s at rse: %s" %
(dataset, destRSE))
if self.dry_run:
print(' Dry run only. Not creating rules')
return
if destRSE:
try:
self.rulesClient.add_replication_rule([{
"scope": self.scope,
"name": "/" + dataset
}],
copies=1,
rse_expression=destRSE)
except DuplicateRule:
print('Rule already exists')
class Rucio :
def __init__(self, myscope, orgRse, destRse, account='bruzzese', working_folder=None):
self.myscope = myscope
self.orgRse = orgRse
self.destRse = destRse
self.working_folder = working_folder
self.gfal = Gfal2Context()
self.didc = DIDClient()
self.repc = ReplicaClient()
self.rulesClient = RuleClient()
# Configuration
self.account = account
# account=account
self.client = Client(account=self.account)
# Get list of all RSEs
def rses(self) :
rses_lists = list()
for single_rse in list(self.client.list_rses()) :
rses_lists.append(single_rse['rse'])
return(rses_lists)
def usage(self,s_rse) :
return(list(self.client.get_local_account_usage(account=self.account,rse=s_rse))[0])
def rules(self) :
return(list(self.client.list_account_rules(account=self.account)))
def myfunc(self):
print("Hello your setting are account=%s, scope=%s, origin RSE =%s and destination RSE =%s" %(self.account, self.myscope, self.orgRse, self.destRse))
def file_exists(self, pfn) :
try :
self.gfal.stat(pfn).st_size
return(True)
except :
return(False)
def get_rse_url(self):
"""
Return the base path of the rucio url
"""
rse_settings = rsemgr.get_rse_info(self.orgRse)
protocol = rse_settings['protocols'][0]
schema = protocol['scheme']
prefix = protocol['prefix']
port = protocol['port']
rucioserver = protocol['hostname']
rse_url = list()
if None not in (schema,str(rucioserver+':'+str(port)),prefix):
rse_url.extend([schema,rucioserver+':'+str(port),prefix,'',''])
if self.working_folder != None :
# Check if our test folder exists
path = os.path.join(urlunsplit(rse_url), self.working_folder)
self.gfal.mkdir_rec(path, 775)
return(path)
else :
return(urlunsplit(rse_url))
else :
return('Wrong url parameters')
def check_replica(self, lfn, dest_rse=None):
"""
Check if a replica of the given file at the site already exists.
"""
if lfn :
replicas = list(
self.client.list_replicas([{
'scope': self.myscope,
'name': lfn
}], rse_expression=dest_rse))
if replicas:
for replica in replicas:
if isinstance(replica,dict) :
if dest_rse in replica['rses']:
path = replica['rses'][dest_rse][0]
return(path)
return(False)
############################
## Create Metadata for DIDs
############################
def getFileMetaData(self, p_file, origenrse=None):
"""
Get the size and checksum for every file in the run from defined path
"""
'''
generate the registration of the file in a RSE :
:param rse: the RSE name.
:param scope: The scope of the file.
:param name: The name of the file.
:param bytes: The size in bytes.
:param adler32: adler32 checksum.
:param pfn: PFN of the file for non deterministic RSE
:param dsn: is the dataset name.
'''
name = os.path.basename(p_file)
name = name.replace('/','')
replica = {
'scope': self.myscope,
'name': name.replace('+','_'),
'adler32': self.gfal.checksum(p_file, 'adler32'),
'bytes': self.gfal.stat(p_file).st_size,
'pfn': p_file,
"meta": {"guid": str(generate_uuid())}
}
Data = dict();
Data['replica'] = replica
Data['scope'] = self.myscope
return(Data)
############################
## Create Groups of DIDs
############################
def createDataset(self, new_dataset) :
logger.debug("| - - Checking if a provided dataset exists: %s for a scope %s" % (new_dataset, self.myscope))
try:
self.client.add_dataset(scope=self.myscope, name=new_dataset)
return(True)
except DataIdentifierAlreadyExists:
return(False)
except Duplicate as error:
return generate_http_error_flask(409, 'Duplicate', error.args[0])
except AccountNotFound as error:
return generate_http_error_flask(404, 'AccountNotFound', error.args[0])
except RucioException as error:
exc_type, exc_obj, tb = sys.exc_info()
logger.debug(exc_obj)
def createcontainer(self, name_container):
'''
registration of the dataset into a container :
:param name_container: the container's name
:param info_dataset : contains,
the scope: The scope of the file.
the name: The dataset name.
'''
logger.debug("| - - - registering container %s" % name_container)
try:
self.client.add_container(scope=self.myscope, name=name_container)
except DataIdentifierAlreadyExists:
logger.debug("| - - - Container %s already exists" % name_container)
except Duplicate as error:
return generate_http_error_flask(409, 'Duplicate', error.args[0])
except AccountNotFound as error:
return generate_http_error_flask(404, 'AccountNotFound', error.args[0])
except RucioException as error:
exc_type, exc_obj, tb = sys.exc_info()
logger.debug(exc_obj)
############################
## General funciotn for registering a did into a GROUP of DID (CONTAINER/DATASET)
############################
def registerIntoGroup(self,n_file, new_dataset):
"""
Attaching a DID to a GROUP
"""
type_1 = self.client.get_did(scope=self.myscope, name=new_dataset)
type_2 = self.client.get_did(scope=self.myscope, name=n_file)
try:
self.client.attach_dids(scope=self.myscope, name=new_dataset, dids=[{'scope':self.myscope, 'name':n_file}])
except RucioException:
logger.debug("| - - - %s already attached to %s" %(type_2['type'],type_1['type']))
############################
## MAGIC functions
############################
def create_groups(self, organization) :
# 2.1) Create the dataset and containers for the file
self.createDataset(organization['dataset_1'])
# 2.1.1) Attach the dataset and containers for the file
self.registerIntoGroup(organization['replica'], organization['dataset_1'])
# 2.2) Create the dataset and containers for the file
self.createcontainer(organization['container_1'])
# 2.2.1) Attach the dataset and containers for the file
self.registerIntoGroup(organization['dataset_1'], organization['container_1'])
# 2.3) Create the dataset and containers for the file
self.createcontainer(organization['container_2'])
# 2.3.1) Attach the dataset and containers for the file
self.registerIntoGroup(organization['container_1'], organization['container_2'])
# 2.4) Create the dataset and containers for the file
self.createcontainer(organization['container_3'])
# 2.4.1) Attach the dataset and containers for the file
self.registerIntoGroup(organization['container_2'], organization['container_3'])
############################
## Create Rule for DIDs
############################
def addReplicaRule(self, destRSE, group):
"""
Create a replication rule for one dataset at a destination RSE
"""
type_1 = self.client.get_did(scope=self.myscope, name=group)
logger.debug("| - - - Creating replica rule for %s %s at rse: %s" % (type_1['type'], group, destRSE))
if destRSE:
try:
rule = self.rulesClient.add_replication_rule([{"scope":self.myscope,"name":group}],copies=1, rse_expression=destRSE, grouping='ALL', account=self.account, purge_replicas=True)
logger.debug("| - - - - Rule succesfully replicated at %s" % destRSE)
logger.debug("| - - - - - The %s has the following id %s" % (rule, destRSE))
return(rule[0])
except DuplicateRule:
exc_type, exc_obj, tb = sys.exc_info()
rules = list(self.client.list_account_rules(account=self.account))
if rules :
for rule in rules :
if rule['rse_expression'] == destRSE and rule['scope'] == self.myscope and rule['name'] == group:
logger.debug('| - - - - Rule already exists %s which contains the following DID %s:%s %s' % (rule['id'],self.myscope, group, str(exc_obj)))
except ReplicationRuleCreationTemporaryFailed:
exc_type, exc_obj, tb = sys.exc_info()
rules = list(self.client.list_account_rules(account=self.account))
if rules :
for rule in rules :
if rule['rse_expression'] == destRSE and rule['scope'] == self.myscope and rule['name'] == group:
print('| - - - - Rule already exists %s which contains the following DID %s:%s %s' % (rule['id'],self.myscope, group, str(exc_obj)))
############################
## Create Rules for not registered DIDs
############################
def outdated_register_replica(self, filemds, dest_RSE, org_RSE):
"""
Register file replica.
"""
carrier_dataset = 'outdated_replication_dataset' + '-' + str(uuid.uuid4())
creation = self.createDataset(carrier_dataset)
# Make sure your dataset is ephemeral
self.client.set_metadata(scope=self.myscope, name=carrier_dataset, key='lifetime', value=86400) # 86400 in seconds = 1 day
# Create a completly new create the RULE:
for filemd in filemds :
outdated = filemd['replica']['name']
self.registerIntoGroup(outdated, carrier_dataset)
# Add dummy dataset for replicating at Destination RSE
rule_child = self.addReplicaRule(dest_RSE, group=carrier_dataset)
# Add dummy dataset for replicating Origin RSE
rule_parent = self.addReplicaRule(org_RSE, group=carrier_dataset)
# Create a relation rule between origin and destiny RSE, so that the source data can be deleted
rule = self.client.update_replication_rule(rule_id=rule_parent, options={'lifetime': 10, 'child_rule_id':rule_child, 'purge_replicas':True})
logger.debug('| - - - - Creating relationship between parent %s and child %s : %s' % (rule_parent, rule_child, rule))
# Create a relation rule between the destinity rule RSE with itself, to delete the dummy rule, whiles keeping the destiny files
rule = self.client.update_replication_rule(rule_id=rule_child, options={'lifetime': 10, 'child_rule_id':rule_child})
logger.debug('| - - - - Creating relationship between parent %s and child %s : %s' % (rule_parent, rule_child, rule))
############################
## Create Dictionary for Grafana
############################
def stats_rules(self, rules) :
'''
Gather general information about
total number of rules, and stats.
'''
RUCIO = dict()
if rules :
for rule in rules :
if 'outdated_replication_dataset' not in rule['name'] :
if 'Rules' not in RUCIO :
RUCIO['Rules'] = {
'total_stuck' : 0,
'total_replicating' : 0,
'total_ok' : 0,
'total_rules': 0
}
RUCIO['Rules']['total_rules'] += 1
if rule['state'] == 'REPLICATING' :
RUCIO['Rules']['total_replicating'] += 1
elif rule['state'] == 'STUCK' :
RUCIO['Rules']['total_stuck'] += 1
elif rule['state'] == 'OK' :
RUCIO['Rules']['total_ok'] += 1
else :
RUCIO['Rules']['total_rules'] += 1
if rule['state'] == 'REPLICATING' :
RUCIO['Rules']['total_replicating'] += 1
elif rule['state'] == 'STUCK' :
RUCIO['Rules']['total_stuck'] += 1
elif rule['state'] == 'OK' :
RUCIO['Rules']['total_ok'] += 1
if 'AllRules' not in RUCIO :
RUCIO['AllRules'] = {
'total_stuck' : 0,
'total_replicating' : 0,
'total_ok' : 0,
'total_rules': 0
}
RUCIO['AllRules']['total_rules'] += 1
if rule['state'] == 'REPLICATING' :
RUCIO['AllRules']['total_replicating'] += 1
elif rule['state'] == 'STUCK' :
RUCIO['AllRules']['total_stuck'] += 1
elif rule['state'] == 'OK' :
RUCIO['AllRules']['total_ok'] += 1
else :
RUCIO['AllRules']['total_rules'] += 1
if rule['state'] == 'REPLICATING' :
RUCIO['AllRules']['total_replicating'] += 1
elif rule['state'] == 'STUCK' :
RUCIO['AllRules']['total_stuck'] += 1
elif rule['state'] == 'OK' :
RUCIO['AllRules']['total_ok'] += 1
##################
if 'Grouping' not in RUCIO :
RUCIO['Grouping'] = {
'file' : 0,
'dataset' : 0,
'container' : 0
}
if rule['did_type'] == 'CONTAINER' :
RUCIO['Grouping']['container'] += 1
elif rule['did_type'] == 'DATASET' :
RUCIO['Grouping']['dataset'] += 1
elif rule['did_type'] == 'FILE' :
RUCIO['Grouping']['file'] += 1
else :
if rule['did_type'] == 'CONTAINER' :
RUCIO['Grouping']['container'] += 1
elif rule['did_type'] == 'DATASET' :
RUCIO['Grouping']['dataset'] += 1
elif rule['did_type'] == 'FILE' :
RUCIO['Grouping']['file'] += 1
return(RUCIO)
def stats_replica_rules(self, rules) :
'''
Gather specific information about
state and number of replicas.
'''
REPLICAS = dict()
REPLICAS['RSE'] = {}
if rules :
# Creates a key for all the RSEs that we have replicas
for rule in rules :
# if the RSE is not in the dictionary
#print(rule['rse_expression'], REPLICAS['RSE'])
if rule['rse_expression'] not in REPLICAS['RSE'] :
#print(REPLICAS)
REPLICAS['RSE'][rule['rse_expression']] = {
'total_replica_stuck' : rule['locks_stuck_cnt'],
'total_replica_replicating' : rule['locks_replicating_cnt'],
'total_replica_ok' : rule['locks_ok_cnt']
}
# else if it is, update replica numbers
else :
REPLICAS['RSE'][rule['rse_expression']]['total_replica_stuck'] += rule['locks_stuck_cnt']
REPLICAS['RSE'][rule['rse_expression']]['total_replica_replicating'] += rule['locks_replicating_cnt']
REPLICAS['RSE'][rule['rse_expression']]['total_replica_ok'] += rule['locks_ok_cnt']
return(REPLICAS)
def stats_usage_rules(self, all_rses) :
STORAGE = dict()
STORAGE['USAGE'] = {}
for x_rse in all_rses :
rses = self.usage(x_rse)
if rses['bytes'] != 0 :
if rses['rse'] not in STORAGE['USAGE'] :
STORAGE['USAGE'][rses['rse']] = {
'total_bytes_used' : rses['bytes']
}
# else if it is, update replica numbers
else :
STORAGE['USAGE'][rses['rse']]['total_bytes_used'] += rses['bytes']
return(STORAGE)