feature(cassandra-stress): use ScyllaDB fork of Cassandra Stress instead of downloading Apache Cassandra

CodeLieutenant · CodeLieutenant · commit 79046fd9c1de · 2025-03-17T14:35:47.000+01:00
Signed-off-by: Dusan Malusev &lt;dusan.malusev@scylladb.com&gt;
diff --git a/.dockerignore b/.dockerignore
@@ -0,0 +1,3 @@
+cassandra-download/
+target/
+.github/
diff --git a/.github/workflows/dockerhub-description.yml b/.github/workflows/dockerhub-description.yml
@@ -18,6 +18,6 @@ jobs:
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
-          repository: scylladb/latte
+          repository: scylladb/cql-stress
           short-description: ${{ github.event.repository.description }}
           enable-url-completion: true
diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml
@@ -5,6 +5,11 @@ on:
     tags:
       - "*"
 
+env:
+  RUSTFLAGS: "--cfg fetch_extended_version_info"
+  CARGO_TERM_COLOR: always
+  PYTHONUNBUFFERED: "1"
+
 jobs:
   get_tag:
     runs-on: ubuntu-latest
@@ -59,35 +64,49 @@ jobs:
         uses: actions/checkout@v4
       - uses: actions-rust-lang/setup-rust-toolchain@v1
 
-      - name: Build Binary
+      - name: Build CQLStressScyllaBench Binary
         uses: actions-rs/cargo@v1
         with:
           command: build
-          args: --release
+          args: --profile dist --all
 
       - name: Set Binary Name
-        id: set_binary_name
         run: |
-          BIN_NAME="cql-stress-${{ needs.get_tag.outputs.tag }}-${{ matrix.os }}"
-          echo "BIN_NAME=$BIN_NAME" >> $GITHUB_ENV
-          echo "BIN_NAME=$BIN_NAME"
+          BIN_NAME_CS="cql-stress-cassandra-stress-${{ needs.get_tag.outputs.tag }}-${{ matrix.os }}"
+          BIN_NAME_SB="cql-stress-scylla-bench-${{ needs.get_tag.outputs.tag }}-${{ matrix.os }}"
+          echo "BIN_NAME_CS=$BIN_NAME_CS" >> $GITHUB_ENV
+          echo "BIN_NAME_CS=$BIN_NAME_CS"
+          echo "BIN_NAME_SB=$BIN_NAME_SB" >> $GITHUB_ENV
+          echo "BIN_NAME_SB=$BIN_NAME_SB"
+
+          mv ./target/dist/cql-stress-cassandra-stress ./target/dist/$BIN_NAME_CS
+          mv ./target/dist/cql-stress-scylla-bench ./target/dist/$BIN_NAME_SB
         shell: bash
 
       - name: Get Release Upload URL
-        id: get_upload_url
         run: |
-          UPLOAD_URL=$(gh release view ${{ inputs.version }} --json uploadUrl -q '.uploadUrl' | sed 's/{.*//')
-          echo "UPLOAD_URL=${UPLOAD_URL}?name=${{ env.BIN_NAME }}" >> $GITHUB_ENV
+          UPLOAD_URL=$(gh release view ${{ needs.get_tag.outputs.tag }} --json uploadUrl -q '.uploadUrl' | sed 's/{.*//')
+          echo "UPLOAD_URL=${UPLOAD_URL}?name=${{ needs.get_tag.outputs.tag }}" >> $GITHUB_ENV
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
 
       - name: Upload Release Asset
-        id: upload-release-asset-unix
+        id: upload-release-asset-cs
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ env.UPLOAD_URL }}
+          asset_path: ./target/dist/${{ env.BIN_NAME_CS }}
+          asset_name: ${{ env.BIN_NAME_CS }}
+          asset_content_type: application/octet-stream
+      - name: Upload Release Asset
+        id: upload-release-asset-sb
         uses: actions/upload-release-asset@v1
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         with:
           upload_url: ${{ env.UPLOAD_URL }}
-          asset_path: ./target/release/cql-stress
-          asset_name: ${{ env.BIN_NAME }}
+          asset_path: ./target/dist/${{ env.BIN_NAME_SB }}
+          asset_name: ${{ env.BIN_NAME_SB }}
           asset_content_type: application/octet-stream
diff --git a/.github/workflows/rust.yml b/.github/workflows/rust.yml
@@ -52,75 +52,67 @@ jobs:
         uses: actions-rs/cargo@v1
         with:
           command: build
-          args: --release --bin cql-stress-cassandra-stress
-      - name: Build cql-stress-scylla-bench binary
-        uses: actions-rs/cargo@v1
-        with:
-          command: build
-          args: --release --bin cql-stress-scylla-bench
+          args: --profile dist --all
 
       - uses: actions/upload-artifact@v4
         with:
           name: cql-stress-cassandra-stress
-          path: "./target/release/cql-stress-cassandra-stress"
+          path: "./target/dist/cql-stress-cassandra-stress"
           if-no-files-found: error
           retention-days: 1
       - uses: actions/upload-artifact@v4
         with:
           name: cql-stress-scylla-bench
-          path: "./target/release/cql-stress-scylla-bench"
+          path: "./target/dist/cql-stress-scylla-bench"
           if-no-files-found: error
           retention-days: 1
   test:
     runs-on: ubuntu-24.04
-    services:
-      scylladb:
-        image: scylladb/scylla
-        ports:
-          - 9042:9042
-        options: >-
-          --name scylla-ci
-          --health-cmd "cqlsh --debug"
-          --health-interval 5s
-          --health-retries 10
-        volumes:
-          - ${{ github.workspace }}:/workspace
     steps:
     - uses: actions/checkout@v4
     - uses: actions-rust-lang/setup-rust-toolchain@v1
       with:
         cache: true
+    - name: Start ScyllaDB
+      run: |
+        docker compose -f ./docker/scylla-test/compose.yml up -d --wait
+        until docker logs scylla_test 2>&1 | grep "Starting listening for CQL clients" > /dev/null; do sleep 0.2; done
     - uses: actions-rs/cargo@v1
       with:
         command: build
-        args: --release --tests --features "user-profile"
+        args: --tests --features "user-profile"
     - uses: actions-rs/cargo@v1
       with:
         command: test
         # test threads must be one because else database tests will run in parallel and will result in flaky tests
         args: --features "user-profile" -- --test-threads=1
 
   integration-tests:
-    needs: [build]
+    needs: [build, test]
+    cancel-timeout-minutes: 1
+    timeout-minutes: 20
     runs-on: ubuntu-24.04
-    services:
-      scylladb:
-        image: scylladb/scylla
-        ports:
-          - 9042:9042
-        options: >-
-          --name scylla-ci
-          --health-cmd "cqlsh --debug"
-          --health-interval 5s
-          --health-retries 10
-        volumes:
-          - ${{ github.workspace }}:/workspace
+    concurrency:
+      group: ${{ github.workflow }}-${{ github.ref }}-${{ matrix.binary }}
+    strategy:
+      max-parallel: 2
+      matrix:
+        binary: [cql-stress-cassandra-stress]
     steps:
       - uses: actions/checkout@v4
-      - uses: actions/download-artifact@v4
       - name: Install python driver
         run: pip install scylla-driver
       - name: Install pytest
         run: pip install -U pytest
+      - name: Start ScyllaDB
+        run: |
+          docker compose -f ./docker/scylla-test/compose.yml up -d --wait
+          until docker logs scylla_test 2>&1 | grep "Starting listening for CQL clients" > /dev/null; do sleep 0.2; done
+      - uses: actions/download-artifact@v4
+        with:
+          name: ${{ matrix.binary }}
       - name: C-S frontend tests
-        run: pytest -s ./tools/cassandra_stress_ci.py
+        run: |
+          mv ${{ matrix.binary }} /usr/local/bin
+          chmod +x /usr/local/bin/${{ matrix.binary }}
+          pytest -s ./tools/${{ matrix.binary }}-ci.py
diff --git a/Cargo.toml b/Cargo.toml
@@ -6,6 +6,31 @@ rust-version = "1.85.0"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
+[[bin]]
+name = "cql-stress-cassandra-stress"
+path = "src/bin/cql-stress-cassandra-stress/main.rs"
+
+[[bin]]
+name = "cql-stress-scylla-bench"
+path = "src/bin/cql-stress-scylla-bench/main.rs"
+
+[profile.release]
+codegen-units = 1
+lto = true
+panic = "abort"
+
+[profile.dist]
+inherits = "release"
+opt-level = 3
+overflow-checks = false
+debug = false
+strip = true
+debug-assertions = false
+
+[profile.dev-opt]
+inherits = "dev"
+opt-level = 2
+
 [dependencies]
 anyhow = "1.0.52"
 async-trait = "0.1.52"
@@ -62,4 +87,6 @@ scylla = { version = "1.0.0", features = [
 ] }
 
 [lints.rust]
-unexpected_cfgs = { level = "warn", check-cfg = ['cfg(fetch_extended_version_info)'] }
+unexpected_cfgs = { level = "warn", check-cfg = [
+    'cfg(fetch_extended_version_info)',
+] }
diff --git a/Dockerfile b/Dockerfile
@@ -0,0 +1,39 @@
+FROM rust:1.85-slim-bookworm AS builder
+
+WORKDIR /app
+
+ENV RUSTFLAGS="--cfg fetch_extended_version_info"
+ENV CARGO_TERM_COLOR=always
+
+COPY . .
+
+RUN apt-get update && apt-get install -y \
+    build-essential \
+    cmake \
+    libclang-dev \
+    git \
+    libssl-dev \
+    pkg-config \
+    && cargo build --profile dist --all
+
+FROM debian:bookworm-slim AS production
+
+ENV PATH="${PATH}:/usr/local/bin"
+
+LABEL org.opencontainers.image.source="https://github.com/scylladb/cql-stress"
+LABEL org.opencontainers.image.title="ScyllaDB cql-stress"
+
+COPY --from=builder /app/target/dist/cql-stress-cassandra-stress /usr/local/bin/cql-stress-cassandra-stress
+COPY --from=builder /app/target/dist/cql-stress-scylla-bench /usr/local/bin/cql-stress-scylla-bench
+
+RUN --mount=type=cache,target=/var/cache/apt apt-get update \
+    && apt-get upgrade -y \
+    && apt-get install -y libssl3 \
+    && apt-get autoremove -y \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* \
+    && chmod +x /usr/local/bin/cql-stress-cassandra-stress /usr/local/bin/cql-stress-scylla-bench \
+    && ln -s /usr/local/bin/cql-stress-cassandra-stress /usr/local/bin/cassandra-stress \
+    && ln -s /usr/local/bin/cql-stress-scylla-bench /usr/local/bin/scylla-bench
+
+ENTRYPOINT [ "bash" ]
diff --git a/docker/scylla-test/compose.yml b/docker/scylla-test/compose.yml
@@ -4,6 +4,7 @@ services:
     container_name: scylla_test
     ports:
       - 9042:9042
+      - 19042:19042
     healthcheck:
       test: [ "CMD", "cqlsh", "-e", "select * from system.local" ]
       interval: 5s
diff --git a/tools/cql-stress-cassandra-stress-ci.py b/tools/cql-stress-cassandra-stress-ci.py
diff --git a/tools/util/cassandra_stress.py b/tools/util/cassandra_stress.py
@@ -6,10 +6,7 @@
 from collections import namedtuple
 
 
-# scylla-tools-java is based on 3.* version of Apache Cassandra.
-# Notice that Apache c-s changed generation logic in version 4.0.0.
-# See: https://github.com/apache/cassandra/commit/f1f5f194620d3f9e11492f0051b6b71018033413
-DEFAULT_CASSANDRA_VERSION = "3.11.19"
+DEFAULT_CASSANDRA_STRESS_VERSION = "v3.17.3"
 ROOT_DIRECTORY = dirname(dirname(dirname(__file__)))
 DOWNLOAD_DIRECTORY_NAME = os.path.join(ROOT_DIRECTORY, "cassandra-download")
 
@@ -79,41 +76,63 @@ def run_user(self, node_ip, profile_name, runtime_args: CSCliRuntimeArguments):
 
 
 class CassandraStress(CSCliRunner):
-    def __init__(self, cassandra_version=DEFAULT_CASSANDRA_VERSION):
-        cassandra_dir = f"apache-cassandra-{cassandra_version}"
-        cassandra_tar = f"{cassandra_dir}-bin.tar.gz"
+    def __init__(self, cassandra_stress_version=DEFAULT_CASSANDRA_STRESS_VERSION):
+        cassandra_stress_dir = "cassandra-stress"
+        cassandra_stress_tar = "cassandra-stress-bin.tar.gz"
 
-        abs_cassandra_dir = os.path.join(
-            DOWNLOAD_DIRECTORY_NAME, cassandra_dir)
-        abs_cassandra_tar = os.path.join(
-            DOWNLOAD_DIRECTORY_NAME, cassandra_tar)
+        abs_cassandra_dir = os.path.join(DOWNLOAD_DIRECTORY_NAME, cassandra_stress_dir)
+        abs_cassandra_tar = os.path.join(DOWNLOAD_DIRECTORY_NAME, cassandra_stress_tar)
+
+        super().__init__(
+            stress_cmd=[os.path.join(abs_cassandra_dir, "bin", "cassandra-stress")]
+        )
 
         if os.path.exists(abs_cassandra_dir):
             # Cassandra already fetched.
             print(
-                f"Cassandra {cassandra_version} already installed. Skipping the download phase.")
-        else:
+                f"Cassandra {cassandra_stress_version} already installed. Skipping the download phase."
+            )
+            return
+
+        if not os.path.exists(abs_cassandra_tar):
+            os.makedirs(DOWNLOAD_DIRECTORY_NAME, exist_ok=True)
             # Fetch cassandra.
             print(
-                f"Fetching cassandra {cassandra_version} to {DOWNLOAD_DIRECTORY_NAME}")
-            # https://dlcdn.apache.org/cassandra/3.11.18/apache-cassandra-3.11.18-bin.tar.gz
-            cassandra_url = f"https://dlcdn.apache.org/cassandra/{cassandra_version}/{cassandra_tar}"
-            subprocess.run(args=["wget", "-P", DOWNLOAD_DIRECTORY_NAME,
-                           "-N", "--no-verbose", cassandra_url], check=True)
-
-            # Extract cassandra
-            print(f"Extracting cassandra {cassandra_version}")
-            subprocess.run(args=["tar", "-xzf", abs_cassandra_tar,
-                           "--directory", DOWNLOAD_DIRECTORY_NAME], check=True)
-            print(f"Extracted cassandra to {abs_cassandra_dir}")
-
-        stress_cmd = [os.path.join(
-            abs_cassandra_dir, "tools", "bin", "cassandra-stress")]
-        super().__init__(stress_cmd=stress_cmd)
+                f"Fetching cassandra {cassandra_stress_version} to {DOWNLOAD_DIRECTORY_NAME}"
+            )
+            cassandra_stress_url = f"https://github.com/scylladb/cassandra-stress/releases/download/{cassandra_stress_version}/cassandra-stress-bin.tar.gz"
+            subprocess.run(
+                args=[
+                    "wget",
+                    "-c",
+                    "-P",
+                    DOWNLOAD_DIRECTORY_NAME,
+                    "-N",
+                    "--no-verbose",
+                    cassandra_stress_url,
+                ],
+                check=True,
+            )
+
+        # Extract cassandra
+        print(f"Extracting cassandra {cassandra_stress_version}")
+        subprocess.run(
+            args=[
+                "tar",
+                "-xzf",
+                abs_cassandra_tar,
+                "--directory",
+                DOWNLOAD_DIRECTORY_NAME,
+            ],
+            check=True,
+        )
+        print(f"Extracted cassandra to {abs_cassandra_dir}")
 
 
 class CqlStressCassandraStress(CSCliRunner):
     def __init__(self):
-        stress_cmd = ["cargo", "run", "--features", "user-profile", "--bin",
-                      "cql-stress-cassandra-stress", "--"]
-        super().__init__(stress_cmd=stress_cmd)
+        super().__init__(stress_cmd=["cql-stress-cassandra-stress"])
+
+
+if __name__ == "__main__":
+    cassandra_stress = CassandraStress()
diff --git a/tools/util/scylla_docker.py b/tools/util/scylla_docker.py
@@ -7,7 +7,7 @@
 
 
 @contextmanager
-def scylla_docker(teardown_behavior, compose_yaml="docker/scylla_test/compose.yml"):
+def scylla_docker(teardown_behavior, compose_yaml="docker/scylla-test/compose.yml"):
     print("Setting up Scylla")
     print("Starting the container and waiting for it to be healthy")
 

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+cassandra-download/`
	`2`	`+target/`
	`3`	`+.github/`