def load_data_in_bucket(self,
folders,
filenames,
missing_field,
operation,
bucket,
start_key=0,
end_key=1000,
batch_size=10,
exp=0,
durability="",
mutation_num=0,
key=None):
"""
Loads data in CB bucket.
:param folders: list, folder paths in aws bucket
:param filenames: list, file names in the bucket
:param missing_field: list of booleans, if missing_field is True,
then this fields value with be omitted while creating a S3 file.
:param operation: create/update/delete
:param bucket: name of the bucket on CB server
:param start_key:
:param end_key:
:param batch_size:
:param exp:
:param durability:
:param mutation_num:
:param key: doc key
:return:
"""
self.log.info("Loading data into bucket")
folder = folders
filename = filenames
missing_field = missing_field
template_obj = JsonObject.create()
template_obj.put("filename", "")
template_obj.put("folder", "")
template_obj.put("mutated", mutation_num)
template_obj.put("null_key", None)
template_obj.put("missing_field", "")
if not key:
key = "test_docs"
doc_gen = DocumentGenerator(key,
template_obj,
start=start_key,
end=end_key,
randomize=True,
filename=filename,
folder=folder,
missing_field=missing_field)
return self.bucket_util.async_load_bucket(self.cluster,
bucket,
doc_gen,
operation,
exp,
durability=durability,
batch_size=batch_size,
suppress_error_table=True)
def load_document_until_ram_percentage(self):
self.start = 0
doc_batch_size = 5000
self.end = doc_batch_size
bucket_helper = BucketHelper(self.cluster.master)
mem_cap = (self.document_ram_percentage * self.bucket_ram * 1000000)
first = ['james', 'sharon', 'dave', 'bill', 'mike', 'steve']
profession = ['doctor', 'lawyer']
template_obj = JsonObject.create()
template_obj.put("number", 0)
template_obj.put("first_name", "")
template_obj.put("profession", "")
template_obj.put("mutated", 0)
template_obj.put("mutation_type", "ADD")
while True:
self.log.info("Add documents to bucket")
doc_gen = DocumentGenerator("test_docs",
template_obj,
start=self.start,
end=self.end,
randomize=False,
first_name=first,
profession=profession,
number=range(70))
try:
self.bucket_util.sync_load_all_buckets(
self.cluster,
doc_gen,
"create",
0,
batch_size=doc_batch_size,
durability=self.durability_level,
suppress_error_table=True)
except Exception as e:
self.fail("Following error occurred while loading bucket - {"
"0}".format(str(e)))
self.log.info("Calculate available free memory")
bucket_json = bucket_helper.get_bucket_json(self.bucket_name)
mem_used = 0
for node_stat in bucket_json["nodes"]:
mem_used += node_stat["interestingStats"]["mem_used"]
if mem_used < mem_cap:
self.log.info("Memory used: %s < %s" % (mem_used, mem_cap))
self.start = self.end
self.end = self.end + doc_batch_size
self.num_items = self.end
else:
break
def perform_doc_ops_in_all_cb_buckets(self, operation,
start_key=0, end_key=1000,
batch_size=10, exp=0,
_async=False,
durability=""):
"""
Create/Update/Delete docs in all cb buckets
:param operation: String - "create","update","delete"
:param start_key: Doc Key to start the operation with
:param end_key: Doc Key to end the operation with
:param batch_size: Batch size of doc_ops
:param exp: MaxTTL used for doc operations
:param _async: Boolean to decide whether to start ops in parallel
:param durability: Durability level to use for doc operation
:return:
"""
first = ['james', 'sharon', 'dave', 'bill', 'mike', 'steve']
profession = ['doctor', 'lawyer']
template_obj = JsonObject.create()
template_obj.put("number", 0)
template_obj.put("first_name", "")
template_obj.put("profession", "")
template_obj.put("mutated", 0)
if operation == "update":
template_obj.put("mutated", 1)
template_obj.put("mutation_type", "ADD")
doc_gen = DocumentGenerator('test_docs', template_obj,
start=start_key, end=end_key,
randomize=True,
first_name=first, profession=profession,
number=range(70))
try:
if _async:
return self.bucket_util._async_load_all_buckets(
self.cluster, doc_gen, operation, exp,
durability=durability,
batch_size=batch_size,
suppress_error_table=True)
else:
self.bucket_util.sync_load_all_buckets(
self.cluster, doc_gen, operation, exp,
durability=durability,
batch_size=batch_size,
suppress_error_table=True)
except Exception as e:
self.log.error(e.message)
def get_doc_generator(self, start, end):
age = range(5)
first = ['james', 'sharon']
body = [''.rjust(self.doc_size - 10, 'a')]
template = JsonObject.create()
template.put("age", None)
template.put("first_name", None)
template.put("body", None)
generator = DocumentGenerator(self.key, template, randomize=True,
age=age,
first_name=first, body=body,
start=start, end=end,
key_size=self.key_size,
doc_size=self.doc_size,
doc_type=self.doc_type)
return generator
def translate_to_json_object(self, value, doc_type="json"):
if type(value) == JsonObject:
return value
json_obj = JsonObject.create()
try:
if doc_type.find("json") != -1:
if type(value) != dict:
value = pyJson.loads(value)
for field, val in value.items():
json_obj.put(field, val)
return json_obj
elif doc_type.find("binary") != -1:
pass
except Exception:
pass
return json_obj
def __load_initial_data(self):
# Create collection meta_data document
sdk_client = self.sdk_clients["bucket_data_writer"]
for tenant in self.tenants:
sdk_client.select_collection(scope_name=tenant,
collection_name="meta_data")
app_data = JsonObject.create()
app_data.put("date", "2001-01-01")
result = sdk_client.crud(DocLoading.Bucket.DocOps.CREATE,
"application", app_data)
self.assertTrue(result["status"], "App_meta creation failed")
self.bucket.scopes[self.tenants[0]].collections["meta_data"]\
.num_items += 1
create_users = User(self.bucket,
scope=tenant,
op_type="scenario_user_registration",
num_items=20000)
create_users.start()
create_users.join()
def get_json_object(document):
"""
Function to convert python dictionary object into Couchbase
JsonObject. This function will recursively convert the dictionary
into JsonObject. Also will add JsonArray as required
:param document: Dictionary to convert to json_object
:type document: dict
:return: Couchbase Jsonobject of the dictionary
:rtype: JsonObject
"""
json_object = JsonObject.create()
for field, val in document.items():
value = None
if isinstance(val, dict):
value = get_json_object(val)
elif isinstance(val, list):
value = get_json_array(val)
else:
value = val
json_object.put(field, value)
return json_object
def perform_doc_ops_in_all_cb_buckets(self,
operation,
start_key=0,
end_key=1000,
batch_size=10,
exp=0,
_async=False,
durability="",
mutation_num=0,
cluster=None,
buckets=[],
key=None):
"""
Create/Update/Delete docs in all cb buckets
:param operation: String - "create","update","delete"
:param start_key: Doc Key to start the operation with
:param end_key: Doc Key to end the operation with
:param batch_size: Batch size of doc_ops
:param exp: MaxTTL used for doc operations
:param _async: Boolean to decide whether to start ops in parallel
:param durability: Durability level to use for doc operation
:param mutation_num: Mutation count to keep track per doc_loading
:param cluster: cluster object for cluster on which this doc load
operation has to be performed.
:param buckets: list of buckets on which doc load operation
has to be performed.
:param key: key for the generated docs
:return:
"""
first = ['james', 'sharon', 'dave', 'bill', 'mike', 'steve']
profession = ['doctor', 'lawyer']
template_obj = JsonObject.create()
template_obj.put("number", 0)
template_obj.put("first_name", "")
template_obj.put("profession", "")
template_obj.put("mutated", mutation_num)
template_obj.put("mutation_type", "ADD")
if not key:
key = "test_docs"
doc_gen = DocumentGenerator(key,
template_obj,
start=start_key,
end=end_key,
randomize=False,
first_name=first,
profession=profession,
number=range(70))
if cluster:
bucket_util = cluster.bucket_util
else:
cluster = self.cluster
bucket_util = self.bucket_util
try:
if _async:
if buckets:
for bucket in buckets:
return bucket_util.async_load_bucket(
cluster,
bucket,
doc_gen,
operation,
exp,
durability=durability,
batch_size=batch_size,
suppress_error_table=True)
else:
return bucket_util._async_load_all_buckets(
cluster,
doc_gen,
operation,
exp,
durability=durability,
batch_size=batch_size,
suppress_error_table=True)
else:
bucket_util.sync_load_all_buckets(cluster,
doc_gen,
operation,
exp,
durability=durability,
batch_size=batch_size,
suppress_error_table=True)
except Exception as e:
self.log.error(e.message)
def get_template():
template = JsonObject.create()
template.put("mutated", 0)
template.put("createdDate", "01/31/1970")
template.put("preferences", JsonObject.create())
return template
def book_flight(u_id, tenant_scope, src_airport=None, dest_airport=None):
summary = dict()
required_seats = choice(range(1, 7))
ticket_type = "normal"
checkout_cart_collection = "checkout_cart"
d_level = Bucket.DurabilityLevel
if [src_airport, dest_airport].count(None) == 0:
ticket_type = "return"
result = query_util.Airline.query_for_routes(
sdk_clients["airline_booking"],
src_airport=src_airport,
dest_airport=dest_airport)
flights = list()
for row in result["q_result"].rowsAsObject():
src_airport = row.get("sourceairport")
dest_airport = row.get("destinationairport")
for flight in row.get("flights"):
flights.append(flight)
summary["src_airport"] = src_airport
summary["dest_airport"] = dest_airport
summary["required_seats"] = required_seats
if not flights:
summary["status"] = "No flights available"
return summary
flight_to_book = choice(flights)
reservation_date = get_random_reservation_date()
checkout_doc = JsonObject.create()
passenger_data = list()
for _ in range(required_seats):
gender = choice(["M", "F"])
first_name = choice(FIRST_NAMES[gender])
last_name = choice(LAST_NAMES)
age = randint(3, 90)
passenger_info = JsonObject.create()
passenger_info.put("first_name", first_name)
passenger_info.put("last_name", last_name)
passenger_info.put("gender", gender)
passenger_info.put("age", age)
passenger_data.append(passenger_info)
client = sdk_clients["airline_booking"]
client.select_collection(tenant_scope, checkout_cart_collection)
cart_id = query_util.CommonUtil.get_next_id(tenant_scope,
checkout_cart_collection)
cart_key = "cart_%s" % cart_id
checkout_doc.put("id", cart_id)
checkout_doc.put("user_id", u_id)
checkout_doc.put("flight_name", flight_to_book.get("flight"))
checkout_doc.put("flight_time", flight_to_book.get("utc"))
checkout_doc.put("travel_date", reservation_date)
checkout_doc.put("from", src_airport)
checkout_doc.put("to", dest_airport)
checkout_doc.put("day_of_week", flight_to_book.get("day"))
checkout_doc.put("seat_count", required_seats)
checkout_doc.put("passengers", passenger_data)
retry = 1
while retry <= User.max_retries:
result = client.crud(DocLoading.Bucket.DocOps.CREATE,
cart_key,
checkout_doc,
durability=d_level.MAJORITY)
if result["status"] is False:
if SDKException.DurabilityImpossibleException \
in str(result["error"]):
User.log.debug("Retrying due to d_impossible")
else:
raise Exception("Flight cart add failed: %s" % result)
retry += 1
if choice([True, False]):
# Booking confirmed scenario, add ticket under flight booking
c_name = "booking_data"
booking_id = query_util.CommonUtil.get_next_id(
tenant_scope, c_name)
ticket_key = "ticket_%s" % booking_id
checkout_doc.put("id", booking_id)
retry = 1
while retry <= User.max_retries:
client.select_collection(tenant_scope, c_name)
result = client.crud(
DocLoading.Bucket.DocOps.CREATE,
ticket_key,
checkout_doc,
durability=d_level.MAJORITY_AND_PERSIST_TO_ACTIVE)
if result["status"] is False:
if SDKException.DurabilityImpossibleException \
in str(result["error"]):
User.log.debug("Retrying due to d_impossible")
else:
raise Exception("Ticket booking failed: %s" % result)
retry += 1
# Add confirmed ticket under user profile
f_booking_id = query_util.CommonUtil.get_next_id(
tenant_scope, "flight_booking")
f_booking_key = "booking_%s" % f_booking_id
f_booking_doc = JsonObject.create()
f_booking_doc.put("id", f_booking_id)
f_booking_doc.put("user_id", u_id)
f_booking_doc.put("ticket_id", booking_id)
f_booking_doc.put("status", "active")
f_booking_doc.put("booked_on", global_vars.app_current_date)
f_booking_doc.put("ticket_type", ticket_type)
client.select_collection(tenant_scope, "flight_booking")
result = client.crud(DocLoading.Bucket.DocOps.CREATE,
f_booking_key, f_booking_doc)
if result["status"] is False:
raise Exception("User flight_booking add failed: %s" % result)
# Remove booked ticket from cart
retry = 1
while retry <= User.max_retries:
client.select_collection(tenant_scope,
checkout_cart_collection)
result = client.crud(DocLoading.Bucket.DocOps.DELETE,
cart_key,
durability=d_level.MAJORITY)
if result["status"] is False:
if SDKException.DurabilityImpossibleException \
in str(result["error"]):
User.log.debug("Retrying due to d_impossible")
else:
break
retry += 1
summary["status"] = "Booking success"
else:
summary["status"] = "cancelled"
return summary
def test_ingestion_after_kv_rollback(self):
try:
for kv_node in self.cluster.kv_nodes:
for bucket in self.cluster.buckets:
self.log.info("Stopping persistence on {0} for bucket {1}"
"".format(kv_node.ip, bucket.name))
mem_client = MemcachedClientHelper.direct_client(
kv_node, bucket)
mem_client.stop_persistence()
# Perform Create, Update, Delete ops in the CB bucket
self.log.info("Performing Mutations")
first = ['james', 'sharon', 'dave', 'bill', 'mike', 'steve']
profession = ['doctor', 'lawyer']
template_obj = JsonObject.create()
template_obj.put("number", 0)
template_obj.put("first_name", "")
template_obj.put("profession", "")
template_obj.put("mutated", 0)
template_obj.put("mutation_type", "ADD")
doc_gen = DocumentGenerator("test_docs",
template_obj,
start=0,
end=1000,
randomize=False,
first_name=first,
profession=profession,
number=range(70))
try:
self.bucket_util.sync_load_all_buckets(
self.cluster,
doc_gen,
"create",
0,
batch_size=1000,
durability=self.durability_level,
suppress_error_table=True)
except Exception as e:
self.fail("Following error occurred while loading bucket - {"
"0}".format(str(e)))
datasets = list()
for dataset in self.cbas_util.list_all_dataset_objs():
if dataset.kv_scope and dataset.kv_scope.name == \
"_default" and dataset.kv_collection.name\
== "_default":
dataset.num_of_items = dataset.num_of_items + 1000
datasets.append(dataset)
for dataset in datasets:
if not self.cbas_util.validate_cbas_dataset_items_count(
self.cluster, dataset.full_name, dataset.num_of_items):
raise Exception(
"Dataset doc count does not match the "
"actual doc count in associated KV collection")
# Kill memcached on Master Node so that another KV node becomes master
self.log.info("Kill Memcached process on Master node")
shell = RemoteMachineShellConnection(self.cluster.master)
shell.kill_memcached()
failover_nodes = list()
# Start persistence on Node
for kv_node in self.cluster.kv_nodes:
if kv_node.ip != self.cluster.master.ip:
failover_nodes.append(kv_node)
for bucket in self.cluster.buckets:
self.log.info(
"Starting persistence on {0} for bucket {1}"
"".format(kv_node.ip, bucket.name))
mem_client = MemcachedClientHelper.direct_client(
kv_node, bucket)
mem_client.start_persistence()
# Failover Nodes on which persistence was started
self.log.info("Failing over Nodes - {0}".format(
str(failover_nodes)))
result = self.task.failover(servers=self.cluster.servers,
failover_nodes=failover_nodes,
graceful=False,
use_hostnames=False,
wait_for_pending=0,
allow_unsafe=False,
all_at_once=False)
self.log.info(str(result))
# Wait for Failover & CBAS rollback to complete
self.sleep(120)
bucket_wise_collection_item_count = dict()
for bucket in self.cluster.buckets:
bucket_wise_collection_item_count[
bucket.
name] = self.bucket_util.get_doc_count_per_collection(
self.cluster, bucket)
self.log.info(str(bucket_wise_collection_item_count))
# Verify the doc count in datasets
for dataset in self.cbas_util.list_all_dataset_objs():
if not self.cbas_util.validate_cbas_dataset_items_count(
self.cluster, dataset.full_name,
bucket_wise_collection_item_count[
dataset.kv_bucket.name][dataset.kv_scope.name][
dataset.kv_collection.name]["items"]):
raise Exception(
"Dataset doc count does not match the actual doc count "
"in associated KV collection after KV roll back")
except Exception as e:
for kv_node in self.cluster.kv_nodes:
for bucket in self.cluster.buckets:
self.log.info("Starting persistence on {0} for bucket {1}"
"".format(kv_node.ip, bucket.name))
mem_client = MemcachedClientHelper.direct_client(
kv_node, bucket)
mem_client.start_persistence()
self.fail(str(e))
def doc_generator(key,
start,
end,
key_size=8,
mix_key_size=False,
doc_size=256,
doc_type="json",
target_vbucket=None,
vbuckets=1024,
mutation_type="ADD",
mutate=0,
randomize_doc_size=False,
randomize_value=False,
randomize=False,
deep_copy=False):
# Defaults to JSON doc_type
template_obj = JsonObject.create()
template_obj.put("mutated", mutate)
_l = len('''{ "mutated": %s
}''' % mutate)
doc_size -= _l
_l = len('"age ": 5,')
if doc_size > _l:
template_obj.put("age", 5)
doc_size -= _l
_l = len('"name ": "james",')
if doc_size > _l:
template_obj.put("name", "james")
doc_size -= _l
_l = len('"mutation_type ": {},'.format(mutation_type))
if doc_size > _l:
template_obj.put("mutation_type", mutation_type)
doc_size -= _l
_l = len('"body ": ')
if doc_size > _l:
template_obj.put("body", "b")
doc_size -= _l
# if doc_type in ["string", "binary"]:
# template_obj = 'age:{0}, first_name: "{1}", body: "{2}", ' \
# 'mutated: %s, mutation_type: "%s"' \
# % (mutate, mutation_type)
if target_vbucket:
return DocumentGeneratorForTargetVbucket(
key,
template_obj,
start=start,
end=end,
key_size=key_size,
mix_key_size=mix_key_size,
doc_size=doc_size,
doc_type=doc_type,
target_vbucket=target_vbucket,
vbuckets=vbuckets,
randomize_doc_size=randomize_doc_size,
randomize_value=randomize_value,
randomize=randomize,
deep_copy=deep_copy)
return DocumentGenerator(key,
template_obj,
start=start,
end=end,
key_size=key_size,
mix_key_size=mix_key_size,
doc_size=doc_size,
doc_type=doc_type,
target_vbucket=target_vbucket,
vbuckets=vbuckets,
randomize_doc_size=randomize_doc_size,
randomize_value=randomize_value,
randomize=randomize,
deep_copy=deep_copy)
def load_initial_collection_data(self):
self.__print_step("Loading initial data into collections")
# Pairs of scope, collection name
type_collection_map = dict()
type_collection_map["airline"] = ("airlines", "airline")
type_collection_map["airport"] = ("airlines", "airport")
type_collection_map["route"] = ("airlines", "routes")
type_collection_map["hotel"] = ("hotels", "hotels")
type_collection_map["landmark"] = ("hotels", "landmark")
meta_data = dict()
for scope in collection_spec["scopes"]:
meta_data[scope["name"]] = dict()
for collection in scope["collections"]:
meta_data[scope["name"]][collection["name"]] = dict()
meta_data[scope["name"]][collection["name"]]["doc_counter"] = 0
meta_data[scope["name"]][collection["name"]]["num_items"] = 0
hotel_review_rows = list()
sdk_client = self.sdk_clients["bucket_data_writer"]
query = "SELECT * FROM `travel-sample`.`_default`.`_default` " \
"WHERE type='%s'"
for d_type, collection_info in type_collection_map.items():
s_name, c_name = collection_info[0], collection_info[1]
sdk_client.select_collection(s_name, c_name)
# TODO: Remove this retry logic once MB-41535 is fixed
retry_index = 0
query_result = None
while retry_index < 5:
try:
query_result = sdk_client.cluster.query(query % d_type)
break
except IndexFailureException:
retry_index += 1
self.sleep(5, "Retrying due to IndexFailure (MB-41535)")
continue
rows_inserted = 0
for row in query_result.rowsAsObject():
value = row.getObject(CbServer.default_collection)\
.removeKey("type")
doc_id = value.getInt("id")
if doc_id > meta_data[s_name][c_name]["doc_counter"]:
meta_data[s_name][c_name]["doc_counter"] = doc_id
if d_type == "hotel":
# Segregate 'reviews' from hotel collection
review = JsonObject.create()
review.put("id", doc_id)
review.put("reviews", value.getArray("reviews"))
hotel_review_rows.append(review)
value = value.removeKey("reviews")
key = d_type + "_" + str(doc_id)
result = sdk_client.crud(DocLoading.Bucket.DocOps.CREATE,
key, value)
if result["status"] is False:
self.fail("Loading collections failed")
rows_inserted += 1
self.bucket.scopes[collection_info[0]].collections[
collection_info[1]].num_items += rows_inserted
meta_data[s_name][c_name]["num_items"] = rows_inserted
# Write hotel reviews into respective collection
rows_inserted = 0
s_name, c_name = "hotels", "reviews"
sdk_client.select_collection(s_name, c_name)
for review in hotel_review_rows:
doc_id = review.getInt("id")
if doc_id > meta_data[s_name][c_name]["doc_counter"]:
meta_data[s_name][c_name]["doc_counter"] = doc_id
key = "review_" + str(doc_id)
result = sdk_client.crud(DocLoading.Bucket.DocOps.CREATE,
key, review)
if result["status"] is False:
self.fail("Loading reviews collection failed")
rows_inserted += 1
self.bucket.scopes[s_name].collections[
c_name].num_items += rows_inserted
meta_data[s_name][c_name]["num_items"] = rows_inserted
# Create collection meta_data document
sdk_client.select_collection(scope_name=CbServer.default_scope,
collection_name="meta_data")
app_data = JsonObject.create()
app_data.put("date", "2001-01-01")
result = sdk_client.crud(DocLoading.Bucket.DocOps.CREATE,
"application", app_data)
self.assertTrue(result["status"], "App_meta creation failed")
self.bucket.scopes[CbServer.default_scope].collections[
"meta_data"].num_items += 1
for s_name, scope_data in meta_data.items():
for c_name, c_data in scope_data.items():
c_meta = JsonObject.create()
for meta_key, meta_val in c_data.items():
c_meta.put(meta_key, meta_val)
key = "%s.%s" % (s_name, c_name)
result = sdk_client.crud(DocLoading.Bucket.DocOps.CREATE,
key, c_meta)
if result["status"] is False:
self.fail("Meta creation failed")
self.bucket.scopes[CbServer.default_scope].collections[
"meta_data"].num_items += 1
create_users = User(self.bucket, op_type="scenario_user_registration",
num_items=10000)
create_users.start()
create_users.join()
self.sleep(30, "Wait for num_items to get updated")
