Update Cassandra DualWrites for using java client

Author: VardhanThigle

sources/cassandra/Dockerfile

@@ -1,9 +1,10 @@
-FROM golang:1.24-bullseye AS builder
+# --- Go Builder Stage ---
+FROM golang:1.24-bullseye AS go-builder
 
 # Disable cgo to remove gcc dependency
 ENV CGO_ENABLED=0
 
-RUN apt-get update && apt-get install -y git curl
+RUN apt-get update && apt-get install -y git
 
 WORKDIR /app
 
@@ -15,30 +16,44 @@ WORKDIR /app/zdm-proxy/proxy
 # Build ZDM Proxy
 RUN go build -o zdm-proxy
 
-# --- Cassandra-to-Spanner CQL Proxy ---
+# --- Java Builder Stage ---
+FROM maven:3.9-eclipse-temurin-17 AS java-builder
+
 WORKDIR /app
 
-RUN git clone "https://github.com/googleapis/go-spanner-cassandra.git"
+# Clone the Java Spanner Cassandra repo
+# Using specific tag or main branch. For stability, usually better to pin, but instructions enable main.
+RUN git clone -b v1.1.0 https://github.com/googleapis/java-spanner-cassandra.git
 
-WORKDIR /app/go-spanner-cassandra
+WORKDIR /app/java-spanner-cassandra
 
-# Build CQL Proxy
-RUN go build -o cassandra-spanner-proxy
+# Build the project to get the shaded jar
+RUN mvn clean install -DskipTests
 
 # --- Final Image ---
 FROM alpine:3.22
 
-# Install bash, ca-certificates, openssl, and update packages
+# Install bash, ca-certificates, openssl, gettext (for envsubst), and Java 17 JRE
 RUN apk update && \
     apk upgrade && \
-    apk add --no-cache bash ca-certificates openssl
+    apk add --no-cache bash ca-certificates openssl gettext openjdk17-jre
 
 EXPOSE 9042 14002
 
-# Copy in the binaries
-COPY --from=builder /app/zdm-proxy/proxy/zdm-proxy .
-COPY --from=builder /app/go-spanner-cassandra/cassandra-spanner-proxy .
+WORKDIR /app
+
+# Copy ZDM Proxy binary
+COPY --from=go-builder /app/zdm-proxy/proxy/zdm-proxy .
+
+# Copy Java Spanner Cassandra Proxy JAR and dependencies
+# The build produces spanner-cassandra-launcher.jar and a lib/ directory with dependencies.
+COPY --from=java-builder /app/java-spanner-cassandra/google-cloud-spanner-cassandra/target/spanner-cassandra-launcher.jar spanner-cassandra-proxy.jar
+COPY --from=java-builder /app/java-spanner-cassandra/google-cloud-spanner-cassandra/target/lib ./lib
+
+# Copy configuration template
+COPY spanner-cassandra-config.yaml .
 
+# Copy entrypoint
 COPY entrypoint.sh /entrypoint.sh
 RUN chmod +x /entrypoint.sh
 

sources/cassandra/README.md

@@ -12,7 +12,7 @@ This repository contains components for doing live migrations from Cassandra to
 - Handles dual writes to both Cassandra and Spanner
 - Ensures data consistency during migration
 
-### 2. Cassandra-Spanner Proxy
+### 2. [Spanner Cassandra Java Client](https://github.com/googleapis/java-spanner-cassandra)
 - Runs as a sidecar to ZDM proxy
 - Translates CQL (Cassandra Query Language) to Spanner API calls
 - Enables Cassandra-compatible applications to interact with Spanner
@@ -26,7 +26,7 @@ The migration setup consists of:
 - **Origin**: Source Cassandra database
 - **Target**: Destination Cloud Spanner database
 - **ZDM Proxy**: Manages dual writes and read routing
-- **Cassandra-Spanner Proxy**: Translates CQL to Spanner API calls
+- **Spanner Cassandra Java Client**: Translates CQL to Spanner API calls
 - **Client Application**: Your application that interacts with the databases
 
 ## Migration Process
@@ -181,7 +181,7 @@ When planning your production deployment, consider these guidelines:
 
 1. **Prepare Terraform Configuration**
 
-   Update your `terraform.tfvars` file with your variables. Take a look at available variables in `variables.tf`. If you need more customization for zdm proxy or cassandra to spanner proxy, update the `variables.tf` and `main.tf` to include the other params.
+   Update your `terraform.tfvars` file with your variables. Take a look at available variables in `variables.tf`. If you need more customization for zdm proxy or java spanner proxy, update the `variables.tf` and `main.tf` to include the other params.
 
 2. **Initialize and Apply Terraform**
 

sources/cassandra/entrypoint.sh

@@ -1,16 +1,29 @@
 #!/bin/bash
 
-# Start processes and store their PIDs
-/cassandra-spanner-proxy --db="projects/$SPANNER_PROJECT/instances/$SPANNER_INSTANCE/databases/$SPANNER_DATABASE" --grpc-channels=500 --tcp="0.0.0.0:9042" &
+# Generated configuration file from template using envsubst
+# envsubst does not support default values in the template (e.g. ${VAR:-default}).
+# We must set the defaults here in the shell before substitution.
+export GRPC_CHANNELS=${GRPC_CHANNELS:-4}
+export MAX_COMMIT_DELAY=${MAX_COMMIT_DELAY:-0}
+
+# Users should ensure SPANNER_PROJECT, SPANNER_INSTANCE, SPANNER_DATABASE are set.
+envsubst < /app/spanner-cassandra-config.yaml > /app/generated-config.yaml
+
+# Start Java Spanner Cassandra Proxy and store PID
+# We use the generated config file.
+java -DconfigFilePath=/app/generated-config.yaml -jar /app/spanner-cassandra-proxy.jar &
 cass_pid=$!
+
 sleep 5
-/zdm-proxy --config="$ZDM_CONFIG" &
+
+# Start ZDM Proxy
+/app/zdm-proxy --config="$ZDM_CONFIG" &
 zdm_pid=$!
 
 # Monitor specific PIDs, exit if either of the processes die.
 while true; do
     if ! kill -0 $cass_pid 2>/dev/null; then
-        echo "cassandra-to-spanner-proxy (PID $cass_pid) died, shutting down container"
+        echo "java-spanner-cassandra-proxy (PID $cass_pid) died, shutting down container"
         exit 1
     fi
     if ! kill -0 $zdm_pid 2>/dev/null; then

sources/cassandra/spanner-cassandra-config.yaml

@@ -0,0 +1,12 @@
+globalClientConfigs:
+  enableBuiltInMetrics: true
+  healthCheckEndpoint: "127.0.0.1:8080"
+
+listeners:
+  - name: "listener_1"
+    host: "127.0.0.1"
+    port: 9042
+    spanner:
+      databaseUri: "projects/${SPANNER_PROJECT}/instances/${SPANNER_INSTANCE}/databases/${SPANNER_DATABASE}"
+      numGrpcChannels: ${GRPC_CHANNELS}
+      maxCommitDelayMillis: ${MAX_COMMIT_DELAY}