def handle(self, *args, **options):
if len(args) == 0:
raise CommandError(
"A dpn_object_id is required to execute this task.")
dpn_obj_id = args[0]
node = Node.objects.get(name=DPN_NODE_NAME)
try:
entry = RegistryEntry.objects.get(dpn_object_id=dpn_obj_id)
if node.name != entry.first_node_name and node not in entry.replicating_nodes.all():
raise CommandError(
"Current node is not listed either as first node or replicating node.")
correlation_id = str(uuid4())
headers = dict(correlation_id=correlation_id)
body = {
"protocol": DPN_XFER_OPTIONS,
"dpn_object_id": dpn_obj_id
}
msg = RecoveryInitQuery(headers, body)
# save initial workflow
action = Workflow(
correlation_id=correlation_id,
dpn_object_id=dpn_obj_id,
action=RECOVERY,
node=DPN_NODE_NAME,
step=INIT_QUERY
)
# handle errors on sending message to network
try:
msg.send(DPN_BROADCAST_KEY)
action.state = SUCCESS
except Exception as err:
action.state = FAILED
action.note = err
# save action in DB
action.save()
# now queue the task responsible to handle the recovery
ttl = DPN_MSG_TTL.get('recovery-init-query', DPN_TTL)
delay = ttl * 2
choose_node_and_recover.apply_async(
(correlation_id, dpn_obj_id, action),
countdown=delay
)
except RegistryEntry.DoesNotExist:
raise CommandError(
"The dpn_object_id that you provided does not exists.")
def on_created(self, event):
if not event.is_directory:
bag_error = False
filename = base = os.path.basename(event.src_path)
initial_filesize = os.path.getsize(event.src_path)
if type(filename) == bytes:
filename = filename.decode("utf-8")
logger.info("New bag detected: %s. Let's wait 5 seconds and check size again..." % base)
while True:
# NOTE: how long should we wait to get the final size of the bag??
# discuss this with the team
# wait 5 seconds to check bag size again
time.sleep(5)
try:
filesize_now = os.path.getsize(event.src_path)
# if initial filesize is equal to the filesize now
# we can start the bag ingestion
if initial_filesize == filesize_now:
filesize = filesize_now
break
else:
initial_filesize = filesize_now
print("Bag is not ready, check again in 5 seconds...")
except Exception as err:
bag_error = err
break
if bag_error:
logger.error("Error processing the new bag %s. Msg -> %s" % (base, bag_error))
elif filesize < DPN_MAX_SIZE:
logger.info("Bag looks good. Starting ingestion of %s..." % base)
# start ingestion and link task to choose nodes
filename_id = os.path.splitext(filename)[0]
delay = DPN_MSG_TTL.get("replication-init-query", DPN_TTL)
initiate_ingest.apply_async(
(filename_id, filesize), link=choose_and_send_location.subtask((), countdown=delay)
)
# execute choose_and_send_location task DPN_TTL seconds after
# ReplicationInitQuery has been sent to broadcast queue
# using countdown parameter of celery task to do that
else:
logger.info("Bag %s is too big to be replicated. Not ingested!" % base)
def handle(self, *args, **options):
start_datetime = self.validate_date(options['startdate'])
end_datetime = self.validate_date(options['enddate'])
if start_datetime > end_datetime:
raise CommandError("Start date must be prior to End Date")
headers = {
'correlation_id': str(uuid4()),
'sequence': 0
}
body = {
'date_range': [dpn_strftime(start_datetime), dpn_strftime(end_datetime)]
}
reg_sync = RegistryDateRangeSync(headers, body)
reg_sync.send(DPN_BROADCAST_KEY)
delay = DPN_MSG_TTL.get("registry-daterange-sync-request", DPN_TTL)
solve_registry_conflicts.apply_async(countdown=delay)
def _get_ttl(self):
return DPN_MSG_TTL.get(self.directive, DPN_TTL)
