Add PostgreSQL + Sysbench TPC-C Workload (#434)

* feat: add postgresql 15

* chore: no default entrypoint for postgre

* feat: Add workloads based on sysbench

* CI: add support for postgresql

* fix: wrong setting for server

* doc: the document draft for postgresql

* fix: spelling

* postgre: docker container entrypoint args python

* postgre: documentation

* fix: missing word

* chore: rename PostgreSQL database

* fix: add missing parameter

---------

Co-authored-by: Rafael Ulises Luzius Pizarro Solar <[email protected]>

.github/workflows/build-images.yaml

@@ -229,6 +229,33 @@ jobs:
       - id: set_is_parent_modified
         run: echo "is_parent_modified=${MODIFIED}" >> $GITHUB_OUTPUT
 
+  postgresql:
+    runs-on: ubuntu-latest
+    needs: base-os
+    env:
+      DH_REPO: "cloudsuite/${{ github.job }}"
+    outputs:
+      is_parent_modified: ${{ steps.set_is_parent_modified.outputs.is_parent_modified }}
+    strategy:
+      matrix:
+        tag: ["15"]
+        platform: ["linux/amd64,linux/arm64"]
+    steps:
+      - name: checkout
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+      - if: ${{ needs.java.outputs.is_parent_modified == 'true' }}
+        run: echo "IS_PARENT_MODIFIED=true" >> $GITHUB_ENV
+      - name: build and push
+        run: "./.github/scripts/build-images.sh"
+        env:
+          IMG_TAG: "${{ matrix.tag }}"
+          DF_PATH: "./commons/${{ github.job }}/${{ matrix.tag }}"
+          DBX_PLATFORM: ${{ matrix.platform }}
+      - id: set_is_parent_modified
+        run: echo "is_parent_modified=${MODIFIED}" >> $GITHUB_OUTPUT
+
   data-analytics:
     runs-on: ubuntu-latest
     needs: hadoop
@@ -301,6 +328,29 @@ jobs:
           DF_PATH: "./benchmarks/${{ github.job }}/${{ matrix.tag }}"
           DBX_PLATFORM: ${{ matrix.platform }}
 
+  data-serving-relational:
+    runs-on: ubuntu-latest
+    needs: postgresql
+    env:
+      DH_REPO: "cloudsuite/${{ github.job }}"
+    strategy:
+      matrix:
+        tag: ["server", "client"]
+        platform: ["linux/amd64,linux/arm64"]
+    steps:
+      - name: checkout
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+      - if: ${{ needs.cassandra.outputs.is_parent_modified == 'true' }}
+        run: echo "IS_PARENT_MODIFIED=true" >> $GITHUB_ENV
+      - name: build and push
+        run: "./.github/scripts/build-images.sh"
+        env:
+          IMG_TAG: "${{ matrix.tag }}"
+          DF_PATH: "./benchmarks/${{ github.job }}/${{ matrix.tag }}"
+          DBX_PLATFORM: ${{ matrix.platform }}
+
   graph-analytics:
     runs-on: ubuntu-latest
     needs: spark

.wordlist.txt

@@ -92,6 +92,7 @@ mysqlserverdocker
 nginx
 NoSQL
 Nutch
+OLTP
 OPERATIONCOUNT
 os
 PageRank
@@ -118,10 +119,14 @@ solr
 solr's
 SQL
 src
+stddev
 sudo
 sys
+sysbench
 taskset
 TCP
+TPC
+tpcc
 threadcount
 txt
 UI

benchmarks/data-serving-relational/client/Dockerfile

@@ -0,0 +1,23 @@
+FROM cloudsuite/base-os:ubuntu
+
+ENV DEBIAN_FRONTEND noninteractive
+
+
+# 1. install necessary software (sysbench)
+RUN apt update && apt install git sysbench python3 -y
+
+# 2. clone sysbench-tpcc's repo and install its script
+RUN git clone https://github.com/Percona-Lab/sysbench-tpcc && cp sysbench-tpcc/*.lua /usr/share/sysbench/
+
+# 3. Copy the template load file
+COPY ./docker-entrypoint.py /root
+COPY ./template/tpcc.py /root/template/tpcc.py
+COPY ./template/oltp-rw.py /root/template/oltp-rw.py
+COPY ./template/database.conf /root/template/database.conf
+ENV DATABASE_CONF_FILE /root/template/database.conf
+
+RUN chmod +x /root/docker-entrypoint.py 
+RUN chmod +x /root/template/tpcc.py
+RUN chmod +x /root/template/oltp-rw.py
+
+ENTRYPOINT ["/root/docker-entrypoint.py"]

benchmarks/data-serving-relational/client/docker-entrypoint.py

@@ -0,0 +1,84 @@
+#!/usr/bin/env python3
+
+import os
+import sys
+import subprocess
+import argparse
+
+args = sys.argv[1:]
+parser = argparse.ArgumentParser()
+parser.add_argument("--tpcc", help="Run TPC-C benchmark", action='store_true')
+parser.add_argument("--oltp-rw", help="Run sysbench OLTP Read/Write workload", action='store_true')
+parser.add_argument("--server-ip", help="IP of the server to load")
+
+args_parsed, unknown = parser.parse_known_args()
+
+if not args_parsed.server_ip:
+    print("Please pass the server IP as an argument with --server-ip=<IP>")
+    sys.exit()
+
+print("args: " + str(args))
+if not args_parsed.tpcc and not args_parsed.oltp_rw:
+    print("Precise whenever it's --tpcc or --oltp-rw")
+    sys.exit()
+
+import os
+import os.path as path
+import shutil
+
+def get_dict(lines):
+    config_dict = {}
+    for line in lines:
+        is_enabled = True 
+        if "=" in line:
+            if line.startswith("#"):
+                is_enabled = False
+                line = line[1:] # Remove `#`
+
+            key, value = line.split("=", 1)
+            key = key.strip()
+            value = value.strip()
+            config_dict[key] = (value, is_enabled)
+
+    return config_dict
+
+def save_dict(config_dict, lines):
+    # Reconstruct the updated configuration
+    new_lines = []
+    for line in lines:
+        if "=" in line:
+            if line.startswith("#"):
+                line = line[1:]
+            key, _ = line.split("=", 1)
+            key = key.strip()
+            if config_dict[key][1]:
+                new_lines.append(f"{key}={config_dict[key][0]}")
+            else:
+                new_lines.append(f"#{key}={config_dict[key][0]}")
+        else:
+            new_lines.append(line)
+
+    new_config = "\n".join(new_lines)
+    return new_config
+
+DATABASE_CONF_FILE = os.environ["DATABASE_CONF_FILE"]
+
+if not path.exists(f"{DATABASE_CONF_FILE}"):
+    shutil.copy(f"{DATABASE_CONF_FILE}", f"{DATABASE_CONF_FILE}.bak")
+
+with open(f"{DATABASE_CONF_FILE}", "r") as f:
+    lines = f.readlines()
+    config_dict = get_dict(lines)
+
+    # Update the desired key with the new value
+    config_dict["pgsql-host"] = (args_parsed.server_ip, True)
+
+    file_txt = save_dict(config_dict, lines)
+    # Write it back
+    with open(f"{DATABASE_CONF_FILE}", "w") as f:
+        f.writelines(file_txt)
+
+if args_parsed.tpcc:
+    subprocess.call(['/root/template/tpcc.py'] + args)
+else:
+    subprocess.call(['/root/template/oltp-rw.py'] + args)

benchmarks/data-serving-relational/client/template/database.conf

@@ -0,0 +1,6 @@
+db-driver=pgsql
+pgsql-host=128.178.116.117
+pgsql-port=5432
+pgsql-user=cloudsuite
+pgsql-password=cloudsuite
+pgsql-db=sbtest

benchmarks/data-serving-relational/client/template/oltp-rw.py

@@ -0,0 +1,40 @@
+#!/usr/bin/env python3
+
+import os
+import sys
+import subprocess
+import argparse
+
+# According to the code, the table structure is like the following:
+# - id: 4B  (primary key)
+# - key: 4B
+# - c: 120B
+# - pad: 60B
+# As a result, each row takes 188B. 
+# You can increase the dataset size by adding more 
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--run", help="Run the benchmark, must be warmuped up before with --warmup", action='store_true')
+parser.add_argument("--warmup", help="Warmup the benchmark, then can be ran with --run", action='store_true')
+parser.add_argument("--threads", "-t", help="Number of threads for the client", default=8, type=int)
+parser.add_argument("--report-interval", "-ri", help="Report interval for metrics in seconds", default=10, type=int)
+parser.add_argument("--record-count", "-c", help="Record count per table. Each record is 188B", default=1000000, type=int)
+parser.add_argument("--tables", "-n", help="Number of tables with `table_size` rows each", default=50, type=int)
+parser.add_argument("--rate", "-r", help="The expected load (transaction / sec)", type=int)
+parser.add_argument("--time", "-s", help="Length of the benchmark in seconds", default=360, type=int)
+
+
+
+args_parsed, unknown = parser.parse_known_args()
+
+# Warmup
+if not args_parsed.warmup and not args_parsed.run:
+    print("Need to pass at least --run or --warmup argument")
+    exit()
+
+if args_parsed.warmup:
+	os.system(f"sysbench oltp_read_write --config-file=/root/template/database.conf --threads={args_parsed.threads} --time={args_parsed.time} --report-interval={args_parsed.report_interval}  prepare --table_size={args_parsed.record_count} --tables={args_parsed.tables}")
+elif not args_parsed.rate:
+	os.system(f"sysbench oltp_read_write --config-file=/root/template/database.conf --threads={args_parsed.threads} --time={args_parsed.time} --report-interval={args_parsed.report_interval} run --table_size={args_parsed.record_count} --tables={args_parsed.tables}")
+else:
+    os.system(f"sysbench oltp_read_write --config-file=/root/template/database.conf --threads={args_parsed.threads} --time={args_parsed.time} --report-interval={args_parsed.report_interval} run --table_size={args_parsed.record_count} --tables={args_parsed.tables} --rate={args_parsed.rate}")

benchmarks/data-serving-relational/client/template/tpcc.py

@@ -0,0 +1,31 @@
+#!/usr/bin/env python3
+
+import os
+import sys
+import subprocess
+import argparse
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--run", help="Run the benchmark, must be warmuped up before with --warmup", action='store_true')
+parser.add_argument("--warmup", help="Warmup the benchmark, then can be ran with --run", action='store_true')
+parser.add_argument("--threads", "-t", help="Number of threads for the client", default=8, type=int)
+parser.add_argument("--report-interval", "-ri", help="Report interval for metrics in seconds", default=10, type=int)
+parser.add_argument("--time", "-s", help="Length of the benchmark in seconds", default=360, type=int)
+parser.add_argument("--scale", "-n", help="Scale of the dataset", default=10, type=int)
+parser.add_argument("--rate", "-r", help="The expected load (transaction / sec)", type=int)
+
+args_parsed, unknown = parser.parse_known_args()
+
+# Warmup
+if not args_parsed.warmup and not args_parsed.run:
+    print("Need to pass at least --run or --warmup argument")
+    exit()
+
+
+
+if args_parsed.warmup:
+	os.system(f"sysbench tpcc --config-file=/root/template/database.conf --threads={args_parsed.threads} prepare --scale={args_parsed.scale}")
+elif not args_parsed.rate:
+	os.system(f"sysbench tpcc --config-file=/root/template/database.conf --threads={args_parsed.threads} --time={args_parsed.time} --report-interval={args_parsed.report_interval} run --scale={args_parsed.scale}")
+else:
+	os.system(f"sysbench tpcc --config-file=/root/template/database.conf --threads={args_parsed.threads} --time={args_parsed.time} --report-interval={args_parsed.report_interval} run --scale={args_parsed.scale} --rate={args_parsed.rate}")

benchmarks/data-serving-relational/server/Dockerfile

@@ -0,0 +1,14 @@
+FROM cloudsuite/postgresql:15
+
+# Install sudo for user switching
+RUN apt update && apt install sudo python3 -y
+
+# Make the database access public
+RUN echo 'host\tall\tcloudsuite\t0.0.0.0/0\tscram-sha-256' >> /etc/postgresql/15/main/pg_hba.conf
+
+# Copy the entrypoint
+COPY ./docker-entrypoint.py /root
+
+RUN chmod +x /root/docker-entrypoint.py
+
+ENTRYPOINT ["/root/docker-entrypoint.py"]

benchmarks/data-serving-relational/server/docker-entrypoint.py

@@ -0,0 +1,88 @@
+#!/usr/bin/env python3
+
+import socket
+def get_ip():
+    s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
+    s.settimeout(0)
+    try:
+        # doesn't even have to be reachable
+        s.connect(('8.8.8.8', 1))
+        IP = s.getsockname()[0]
+    except Exception:
+        IP = '127.0.0.1'
+    finally:
+        s.close()
+    return IP
+
+import argparse
+
+parser = argparse.ArgumentParser()
+# If no value provided, the script tries to find the primary IP address by itself. = get_ip()
+parser.add_argument("--listen-addresses", "-a", help="The listening IP address of PostGRES.", default="'*'")
+parser.add_argument("--number", "-n", type=int, help="The number is not used, place holder for new argument.", default=0)
+
+args, unknown = parser.parse_known_args()
+
+import os
+import os.path as path
+import shutil
+
+def get_dict(lines):
+    config_dict = {}
+    for line in lines:
+        is_enabled = True 
+        if "=" in line:
+            if line.startswith("#"):
+                is_enabled = False
+                line = line[1:] # Remove `#`
+
+            key, value = line.split("=", 1)
+            key = key.strip()
+            value = value.strip()
+            config_dict[key] = (value, is_enabled)
+
+    return config_dict
+
+def save_dict(config_dict, lines):
+    # Reconstruct the updated configuration
+    new_lines = []
+    for line in lines:
+        if "=" in line:
+            if line.startswith("#"):
+                line = line[1:]
+            key, _ = line.split("=", 1)
+            key = key.strip()
+            if config_dict[key][1]:
+                new_lines.append(f"{key} = {config_dict[key][0]}")
+            else:
+                new_lines.append(f"#{key} = {config_dict[key][0]}")
+        else:
+            new_lines.append(line)
+
+    new_config = "\n".join(new_lines)
+    return new_config
+
+POSTGRE_HOMEDIR = os.environ["POSTGRE_HOME"]
+
+# Backup the original file
+if not path.exists(f"{POSTGRE_HOMEDIR}/postgresql.conf"):
+    shutil.copy(f"{POSTGRE_HOMEDIR}/postgresql.conf", f"{POSTGRE_HOMEDIR}/postgresql.conf.bak")
+
+with open(f"{POSTGRE_HOMEDIR}/postgresql.conf", "r") as f:
+    lines = f.readlines()
+    config_dict = get_dict(lines)
+
+    # Update the desired key with the new value
+    config_dict["listen_addresses"] = (args.listen_addresses, True) # sed -i "s/#listen_addresses = 'localhost'/listen_addresses = '*'/g" /etc/postgresql/15/main/postgresql.conf
+
+    file_txt = save_dict(config_dict, lines)
+    # Write it back
+    with open(f"{POSTGRE_HOMEDIR}/postgresql.conf", "w") as f:
+        f.writelines(file_txt)
+
+os.system("service postgresql start")
+os.system("sudo -u postgres psql -c \"CREATE USER cloudsuite WITH PASSWORD 'cloudsuite';\"") # Create the user called `cloudsuite`
+os.system("sudo -u postgres psql -c \"CREATE DATABASE sbtest;\"") # Create a table named sbtest
+os.system("sudo -u postgres psql -c \"GRANT ALL PRIVILEGES ON DATABASE sbtest TO cloudsuite\"") # Gave permission to this table
+os.system("sudo -u postgres psql sbtest -c \"GRANT ALL ON SCHEMA public TO cloudsuite;\"")
+os.system("sudo -u postgres psql")

commons/postgresql/15/Dockerfile

@@ -0,0 +1,12 @@
+FROM cloudsuite/base-os:ubuntu
+
+ENV DEBIAN_FRONTEND noninteractive
+
+RUN apt update && apt install -y wget gnupg lsb-release && \
+    echo "deb http://apt.postgresql.org/pub/repos/apt $(lsb_release -cs)-pgdg main" > /etc/apt/sources.list.d/pgdg.list && \
+    wget --quiet -O - https://www.postgresql.org/media/keys/ACCC4CF8.asc | apt-key add - && \
+    apt update && \
+    apt -y install postgresql-15
+
+
+ENV POSTGRE_HOME /etc/postgresql/15/main