V0.5.14 (#92)

* Prepare next release

* TPC-DS: Updated query files

* Evaluation: Enable toc in jupyter

* TPC-DS: Updated query files and title

* TPC-DS: Updated example

* TPC-H: Improved experiment management

* TPC-H: Improved experiment management - example file

* TPC-H: Improved experiment management

* TPC-H: Improved examples

* PVC for MonetDB

bexhoma/experiments.py

@@ -375,7 +375,7 @@ def stop_monitoring(self):
 	def start_monitoring(self):
 		for config in self.configurations:
 			config.start_monitoring()
-	def start_sut(self):
+	def start_sut(self):#, configuration=None):
 		for config in self.configurations:
 			config.start_sut()
 	def stop_sut(self):
@@ -398,7 +398,7 @@ def load_data(self):
 	def add_benchmark_list(self, list_clients):
 		for config in self.configurations:
 			config.add_benchmark_list(list_clients)
-	def work_benchmark_list(self, intervals=30):
+	def work_benchmark_list(self, intervals=30, stop=True):
 		do = True
 		while do:
 			time.sleep(intervals)
@@ -473,25 +473,28 @@ def work_benchmark_list(self, intervals=30):
 							config.run_benchmarker_pod(connection=connection, configuration=config.configuration, client=client, parallelism=parallelism)
 						else:
 							# no list element left
-							print("{} can be stopped".format(config.configuration))
-							app = self.cluster.appname
-							component = 'sut'
-							pods = self.cluster.getPods(app, component, self.code, config.configuration)
-							if len(pods) > 0:
-								pod_sut = pods[0]
-								self.cluster.store_pod_log(pod_sut, 'dbms')
-							config.stop_sut()
-							config.numExperimentsDone = config.numExperimentsDone + 1
-							if config.numExperimentsDone < config.numExperiments:
-								print("{} starts again".format(config.configuration))
-								config.benchmark_list = config.benchmark_list_template.copy()
-								# wait for PV to be gone completely
-								self.wait(60)
-								config.reset_sut()
-								config.start_sut()
-								self.wait(10)
+							if stop:
+								print("{} can be stopped".format(config.configuration))
+								app = self.cluster.appname
+								component = 'sut'
+								pods = self.cluster.getPods(app, component, self.code, config.configuration)
+								if len(pods) > 0:
+									pod_sut = pods[0]
+									self.cluster.store_pod_log(pod_sut, 'dbms')
+								config.stop_sut()
+								config.numExperimentsDone = config.numExperimentsDone + 1
+								if config.numExperimentsDone < config.numExperiments:
+									print("{} starts again".format(config.configuration))
+									config.benchmark_list = config.benchmark_list_template.copy()
+									# wait for PV to be gone completely
+									self.wait(60)
+									config.reset_sut()
+									config.start_sut()
+									self.wait(10)
+								else:
+									config.experiment_done = True
 							else:
-								config.experiment_done = True
+								print("{} can be stopped, be we leave it running".format(config.configuration))
 				else:
 					print("{} is loading".format(config.configuration))
 			# all jobs of configuration - benchmarker
@@ -588,11 +591,10 @@ def benchmark_list(self, list_clients):
 						self.cluster.deleteJob(job)
 				if len(pods) == 0 and len(jobs) == 0:
 					break
-	def stop_benchmarker(self):
+	def stop_benchmarker(self, configuration=''):
 		# all jobs of configuration - benchmarker
 		app = self.appname
 		component = 'benchmarker'
-		configuration = ''
 		jobs = self.cluster.getJobs(app, component, self.code, configuration)
 		# status per job
 		for job in jobs:

bexhoma/masterK8s.py

@@ -1613,12 +1613,10 @@ def stop_sut(self, app='', component='sut', experiment='', configuration=''):
             self.deleteStatefulSet(stateful_set)
         if component == 'sut':
             self.stop_sut(app=app, component='worker', experiment=experiment, configuration=configuration)
-    def stop_benchmarker(self):
+    def stop_benchmarker(self, experiment='', configuration=''):
         # all jobs of configuration - benchmarker
         app = self.appname
         component = 'benchmarker'
-        experiment = ''
-        configuration = ''
         jobs = self.getJobs(app, component, experiment, configuration)
         # status per job
         for job in jobs:

bexhoma/scripts/experimentsmanager.py

@@ -43,13 +43,16 @@ def manage():
 	parser.add_argument('-cx', '--context', help='context of Kubernetes (for a multi cluster environment), default is current context', default=None)
 	clusterconfig = 'cluster.config'
 	args = parser.parse_args()
+	connection = args.connection
 	if args.mode == 'stop':
 		cluster = clusters.kubernetes(clusterconfig, context=args.context)
 		if args.experiment is None:
 			experiment = experiments.default(cluster=cluster, code=cluster.code)
-			cluster.stop_sut()
-			cluster.stop_monitoring()
-			cluster.stop_benchmarker()
+			if connection is None:
+				connection = ''
+			cluster.stop_sut(configuration=connection)
+			cluster.stop_monitoring(configuration=connection)
+			cluster.stop_benchmarker(configuration=connection)
 		else:
 			experiment = experiments.default(cluster=cluster, code=args.experiment)
 			experiment.stop_sut()
@@ -60,7 +63,7 @@ def manage():
 		cluster.connect_dashboard()
 	elif args.mode == 'master':
 		cluster = clusters.kubernetes(clusterconfig, context=args.context)
-		cluster.connect_master(experiment=args.experiment, configuration=args.connection)
+		cluster.connect_master(experiment=args.experiment, configuration=connection)
 	elif args.mode == 'status':
 		cluster = clusters.kubernetes(clusterconfig, context=args.context)
 		app = cluster.appname

bexhoma/scripts/tpcds.py

@@ -85,14 +85,14 @@ def do_benchmark():
 		# we want all TPC-H queries
 		experiment.set_queries_full()
 		experiment.set_workload(
-			name = 'TPC-H Queries SF='+str(SF),
+			name = 'TPC-DS Queries SF='+str(SF),
 			info = 'This experiment compares run time and resource consumption of TPC-DS queries in different DBMS.'
 		)
 	else:
 		# we want to profile the import
 		experiment.set_queries_profiling()
 		experiment.set_workload(
-			name = 'TPC-H Data Profiling SF='+str(SF),
+			name = 'TPC-DS Data Profiling SF='+str(SF),
 			info = 'This experiment compares imported TPC-DS data sets in different DBMS.'
 		)
 	if monitoring:
@@ -158,7 +158,7 @@ def do_benchmark():
 		start_datetime = str(datetime.datetime.now())
 		print("Experiment starts at {} ({})".format(start_datetime, start))
 		# run workflow
-		experiment.work_benchmark_list()
+		experiment.work_benchmark_list(stop=False)
 		# total time of experiment
 		end = default_timer()
 		end_datetime = str(datetime.datetime.now())
@@ -175,7 +175,7 @@ def do_benchmark():
 		start_datetime = str(datetime.datetime.now())
 		print("Experiment starts at {} ({})".format(start_datetime, start))
 		# run workflow
-		experiment.work_benchmark_list()
+		experiment.work_benchmark_list(stop=True)
 		# total time of experiment
 		end = default_timer()
 		end_datetime = str(datetime.datetime.now())

bexhoma/scripts/tpch.py

@@ -11,7 +11,6 @@
 """
 from bexhoma import *
 from dbmsbenchmarker import *
-#import experiments
 import logging
 import urllib3
 import logging
@@ -34,6 +33,7 @@ def do_benchmark():
 	# argparse
 	parser = argparse.ArgumentParser(description=description)
 	parser.add_argument('mode', help='profile the import of TPC-H data, or run the TPC-H queries, or start DBMS and load data, or just start the DBMS', choices=['profiling', 'run', 'start', 'load'])
+	parser.add_argument('-c', '--connection', help='name of DBMS', default=None)
 	parser.add_argument('-cx', '--context', help='context of Kubernetes (for a multi cluster environment), default is current context', default=None)
 	parser.add_argument('-e', '--experiment', help='sets experiment code for continuing started experiment', default=None)
 	parser.add_argument('-d', '--detached', help='puts most of the experiment workflow inside the cluster', action='store_true')
@@ -57,6 +57,7 @@ def do_benchmark():
 	args = parser.parse_args()
 	# set parameter
 	monitoring = args.monitoring
+	connection = args.connection
 	mode = str(args.mode)
 	SF = str(args.scaling_factor)
 	timeout = int(args.timeout)
@@ -145,6 +146,12 @@ def do_benchmark():
 	#config = configurations.default(experiment=experiment, docker='Clickhouse', configuration='Clickhouse-{}'.format(cluster_name), alias='DBMS K')
 	#config = configurations.default(experiment=experiment, docker='SQLServer', configuration='SQLServer-{}'.format(cluster_name), alias='DBMS L')
 	#config = configurations.default(experiment=experiment, docker='OmniSci', configuration='OmniSci-{}'.format(cluster_name), alias='DBMS M')
+	if connection is not None:
+		for c in reversed(range(len(experiment.configurations))):
+			if experiment.configurations[c].docker != connection:
+				del experiment.configurations[c]
+	for c in experiment.configurations:
+		print("System to test:", c.configuration)
 	if args.mode == 'start':
 		experiment.start_sut()
 	elif args.mode == 'load':

experiments/tpcds/queries-tpcds.config

@@ -7762,7 +7762,7 @@
 	},
 	{
 		'title': "TPC-DS Q72",
-		'query': """ select  i_item_desc
+		'query': """select i_item_desc
 			,w_warehouse_name
 			,d1.d_week_seq
 			,sum(case when p_promo_sk is null then 1 else 0 end) no_promo
@@ -7789,7 +7789,7 @@
 			order by total_cnt desc, i_item_desc, w_warehouse_name, d_week_seq
 			limit 100""",
 		'DBMS': {
-			'MonetDB': """ select  i_item_desc
+			'MonetDB': """select i_item_desc
 				,w_warehouse_name
 				,d1.d_week_seq
 				,sum(case when p_promo_sk is null then 1 else 0 end) no_promo
@@ -7808,7 +7808,7 @@
 				left outer join catalog_returns on (cr_item_sk = cs_item_sk and cr_order_number = cs_order_number)
 				where d1.d_week_seq = d2.d_week_seq
 				and inv_quantity_on_hand < cs_quantity 
-				and d3.d_date > (cast(d1.d_date as date) + interval 5 days)
+				and d3.d_date > (cast(d1.d_date as date) + interval '5' day)
 				and hd_buy_potential = '{BP}'
 				and d1.d_year = {YEAR}
 				and cd_marital_status = '{MS}'

experiments/tpch/MySQL/initdata-tpch-SF1.sql

@@ -1,32 +1,32 @@
-LOAD DATA INFILE '/data/tpch/SF1/customer.tbl' IGNORE INTO TABLE tpch.customer
+LOAD DATA INFILE '/tmp/tpch/SF1/customer.tbl' IGNORE INTO TABLE tpch.customer
 FIELDS TERMINATED BY '|'
 (@c_custkey, @c_name, @c_address, @c_nationkey, @c_phone, @c_acctbal, @c_mktsegment, @c_comment) SET c_custkey=NULLIF(@c_custkey,''), c_name=NULLIF(@c_name,''), c_address=NULLIF(@c_address,''), c_nationkey=NULLIF(@c_nationkey,''), c_phone=NULLIF(@c_phone,''), c_acctbal=NULLIF(@c_acctbal,''), c_mktsegment=NULLIF(@c_mktsegment,''), c_comment=NULLIF(@c_comment,'');
 
-LOAD DATA INFILE '/data/tpch/SF1/lineitem.tbl' IGNORE INTO TABLE tpch.lineitem
+LOAD DATA INFILE '/tmp/tpch/SF1/lineitem.tbl' IGNORE INTO TABLE tpch.lineitem
 FIELDS TERMINATED BY '|'
 (@l_orderkey, @l_partkey, @l_suppkey, @l_linenumber, @l_quantity, @l_extendedprice, @l_discount, @l_tax, @l_returnflag, @l_linestatus, @l_shipdate, @l_commitdate, @l_receiptdate, @l_shipinstruct, @l_shipmode, @l_comment) SET l_orderkey=NULLIF(@l_orderkey,''), l_partkey=NULLIF(@l_partkey,''), l_suppkey=NULLIF(@l_suppkey,''), l_linenumber=NULLIF(@l_linenumber,''), l_quantity=NULLIF(@l_quantity,''), l_extendedprice=NULLIF(@l_extendedprice,''), l_discount=NULLIF(@l_discount,''), l_tax=NULLIF(@l_tax,''), l_returnflag=NULLIF(@l_returnflag,''), l_linestatus=NULLIF(@l_linestatus,''), l_shipdate=NULLIF(@l_shipdate,''), l_commitdate=NULLIF(@l_commitdate,''), l_receiptdate=NULLIF(@l_receiptdate,''), l_shipinstruct=NULLIF(@l_shipinstruct,''), l_shipmode=NULLIF(@l_shipmode,''), l_comment=NULLIF(@l_comment,'');
 
-LOAD DATA INFILE '/data/tpch/SF1/nation.tbl' IGNORE INTO TABLE tpch.nation
+LOAD DATA INFILE '/tmp/tpch/SF1/nation.tbl' IGNORE INTO TABLE tpch.nation
 FIELDS TERMINATED BY '|'
 (@n_nationkey, @n_name, @n_regionkey, @n_comment) SET n_nationkey=NULLIF(@n_nationkey,''), n_name=NULLIF(@n_name,''), n_regionkey=NULLIF(@n_regionkey,''), n_comment=NULLIF(@n_comment,'');
 
-LOAD DATA INFILE '/data/tpch/SF1/orders.tbl' IGNORE INTO TABLE tpch.orders
+LOAD DATA INFILE '/tmp/tpch/SF1/orders.tbl' IGNORE INTO TABLE tpch.orders
 FIELDS TERMINATED BY '|'
 (@o_orderkey, @o_custkey, @o_orderstatus, @o_totalprice, @o_orderdate, @o_orderpriority, @o_clerk, @o_shippriority, @o_comment) SET o_orderkey=NULLIF(@o_orderkey,''), o_custkey=NULLIF(@o_custkey,''), o_orderstatus=NULLIF(@o_orderstatus,''), o_totalprice=NULLIF(@o_totalprice,''), o_orderdate=NULLIF(@o_orderdate,''), o_orderpriority=NULLIF(@o_orderpriority,''), o_clerk=NULLIF(@o_clerk,''), o_shippriority=NULLIF(@o_shippriority,''), o_comment=NULLIF(@o_comment,'');
 
-LOAD DATA INFILE '/data/tpch/SF1/part.tbl' IGNORE INTO TABLE tpch.part
+LOAD DATA INFILE '/tmp/tpch/SF1/part.tbl' IGNORE INTO TABLE tpch.part
 FIELDS TERMINATED BY '|'
 (@p_partkey, @p_name, @p_mfgr, @p_brand, @p_type, @p_size, @p_container, @p_retailprice, @p_comment) SET p_partkey=NULLIF(@p_partkey,''), p_name=NULLIF(@p_name,''), p_mfgr=NULLIF(@p_mfgr,''), p_brand=NULLIF(@p_brand,''), p_type=NULLIF(@p_type,''), p_size=NULLIF(@p_size,''), p_container=NULLIF(@p_container,''), p_retailprice=NULLIF(@p_retailprice,''), p_comment=NULLIF(@p_comment,'');
 
-LOAD DATA INFILE '/data/tpch/SF1/partsupp.tbl' IGNORE INTO TABLE tpch.partsupp
+LOAD DATA INFILE '/tmp/tpch/SF1/partsupp.tbl' IGNORE INTO TABLE tpch.partsupp
 FIELDS TERMINATED BY '|'
 (@ps_partkey, @ps_suppkey, @ps_availqty, @ps_supplycost, @ps_comment) SET ps_partkey=NULLIF(@ps_partkey,''), ps_suppkey=NULLIF(@ps_suppkey,''), ps_availqty=NULLIF(@ps_availqty,''), ps_supplycost=NULLIF(@ps_supplycost,''), ps_comment=NULLIF(@ps_comment,'');
 
-LOAD DATA INFILE '/data/tpch/SF1/region.tbl' IGNORE INTO TABLE tpch.region
+LOAD DATA INFILE '/tmp/tpch/SF1/region.tbl' IGNORE INTO TABLE tpch.region
 FIELDS TERMINATED BY '|'
 (@r_regionkey, @r_name, @r_comment) SET r_regionkey=NULLIF(@r_regionkey,''), r_name=NULLIF(@r_name,''), r_comment=NULLIF(@r_comment,'');
 
-LOAD DATA INFILE '/data/tpch/SF1/supplier.tbl' IGNORE INTO TABLE tpch.supplier
+LOAD DATA INFILE '/tmp/tpch/SF1/supplier.tbl' IGNORE INTO TABLE tpch.supplier
 FIELDS TERMINATED BY '|'
 (@s_suppkey, @s_name, @s_address, @s_nationkey, @s_phone, @s_acctbal, @s_comment) SET s_suppkey=NULLIF(@s_suppkey,''), s_name=NULLIF(@s_name,''), s_address=NULLIF(@s_address,''), s_nationkey=NULLIF(@s_nationkey,''), s_phone=NULLIF(@s_phone,''), s_acctbal=NULLIF(@s_acctbal,''), s_comment=NULLIF(@s_comment,'');
 

experiments/tpch/MySQL/loadindexes-tpch.sql

@@ -0,0 +1,17 @@
+SELECT COUNT(*) FROM tpch.customer FORCE INDEX(c_n);
+SELECT COUNT(*) FROM tpch.nation FORCE INDEX(n_r);
+SELECT COUNT(*) FROM tpch.supplier FORCE INDEX(s_n);
+SELECT COUNT(*) FROM tpch.partsupp FORCE INDEX(ps_s);
+SELECT COUNT(*) FROM tpch.partsupp FORCE INDEX(ps_p);
+SELECT COUNT(*) FROM tpch.orders FORCE INDEX(o_c);
+SELECT COUNT(*) FROM tpch.lineitem FORCE INDEX(l_o);
+SELECT COUNT(*) FROM tpch.lineitem FORCE INDEX(l_ps);
+
+SELECT COUNT(*) FROM tpch.customer;
+SELECT COUNT(*) FROM tpch.lineitem;
+SELECT COUNT(*) FROM tpch.nation;
+SELECT COUNT(*) FROM tpch.supplier;
+SELECT COUNT(*) FROM tpch.partsupp;
+SELECT COUNT(*) FROM tpch.orders;
+SELECT COUNT(*) FROM tpch.region;
+SELECT COUNT(*) FROM tpch.part;

experiments/tpch/README.md

@@ -8,6 +8,18 @@ It also starts an evaluation dashboard.
 
 Recommended settings: `tpch run -sf 1 -t 30 -dt`
 
+For more detail about options.
+
+## Tutorials
+
+### Install Data
+
+### Start a DBMS
+
+### Start a DBMS and Load Data
+
+### Start a DBMS and Load Data and Connect
+
 ## Options
 
 ```

images/dashboard/Dockerfile

@@ -16,6 +16,14 @@ RUN pip install dbmsbenchmarker
 
 RUN pip install jupyter
 
+RUN pip install jupyter_contrib_nbextensions
+
+RUN jupyter contrib nbextension install --user
+
+RUN jupyter nbextensions_configurator enable --user
+
+RUN jupyter nbextension enable toc2/main
+
 RUN git clone https://github.com/Beuth-Erdelt/DBMS-Benchmarker
 
 WORKDIR /usr/src/app/DBMS-Benchmarker