feat: support both amd64 and arm64 architectures (#69)

Author: giovtorres

.github/workflows/test.yml

@@ -8,67 +8,29 @@ on:
   workflow_dispatch:  # Allow manual trigger
 
 jobs:
-  # Build images for all supported versions
-  build:
-    runs-on: ubuntu-latest
-    timeout-minutes: 30
-    strategy:
-      matrix:
-        slurm_version: ['25.05.3', '24.11.6']
-
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v3
-
-      - name: Build Docker image for Slurm ${{ matrix.slurm_version }}
-        run: |
-          echo "SLURM_VERSION=${{ matrix.slurm_version }}" > .env
-          docker compose build
-
-      - name: Save Docker image
-        run: |
-          docker save slurm-docker-cluster:${{ matrix.slurm_version }} | gzip > slurm-${{ matrix.slurm_version }}.tar.gz
-
-      - name: Upload image artifact
-        uses: actions/upload-artifact@v4
-        with:
-          name: slurm-image-${{ matrix.slurm_version }}
-          path: slurm-${{ matrix.slurm_version }}.tar.gz
-          retention-days: 1
-
-  # Test cluster functionality on all versions
   test-cluster:
-    needs: build
-    runs-on: ubuntu-latest
+    runs-on: ${{ matrix.runner }}
     timeout-minutes: 30
     strategy:
+      fail-fast: false  # Allow all matrix jobs to complete independently
       matrix:
         slurm_version: ['25.05.3', '24.11.6']
+        runner: ['ubuntu-latest', 'ubuntu-24.04-arm']
+        include:
+          - runner: ubuntu-latest
+            arch: amd64
+          - runner: ubuntu-24.04-arm
+            arch: arm64
 
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
 
-      - name: Download image artifact
-        uses: actions/download-artifact@v4
-        with:
-          name: slurm-image-${{ matrix.slurm_version }}
-
-      - name: Load Docker image
-        run: |
-          docker load < slurm-${{ matrix.slurm_version }}.tar.gz
-          docker images
-
-      - name: Configure version
+      - name: Build and start cluster for Slurm ${{ matrix.slurm_version }} (${{ matrix.arch }})
         run: |
           echo "SLURM_VERSION=${{ matrix.slurm_version }}" > .env
           cat .env
-
-      - name: Start cluster with 2 nodes
-        run: |
+          docker compose build
           docker compose up -d
           echo "Waiting for services to start..."
           sleep 15

Dockerfile

@@ -10,6 +10,7 @@ ARG SLURM_VERSION
 FROM rockylinux/rockylinux:9 AS builder
 
 ARG SLURM_VERSION
+ARG TARGETARCH
 
 # Enable CRB and EPEL repositories for development packages
 RUN set -ex \
@@ -66,12 +67,19 @@ RUN rpmdev-setuptree
 COPY rpmbuild/slurm.rpmmacros /root/.rpmmacros
 
 # Download official Slurm release tarball and build RPMs with slurmrestd enabled
+# Architecture mapping: Docker TARGETARCH (amd64, arm64) -> RPM arch (x86_64, aarch64)
 RUN set -ex \
+    && RPM_ARCH=$(case "${TARGETARCH}" in \
+         amd64) echo "x86_64" ;; \
+         arm64) echo "aarch64" ;; \
+         *) echo "Unsupported architecture: ${TARGETARCH}" && exit 1 ;; \
+       esac) \
+    && echo "Building Slurm RPMs for architecture: ${RPM_ARCH}" \
     && wget -O /root/rpmbuild/SOURCES/slurm-${SLURM_VERSION}.tar.bz2 \
        https://download.schedmd.com/slurm/slurm-${SLURM_VERSION}.tar.bz2 \
     && cd /root/rpmbuild/SOURCES \
     && rpmbuild -ta slurm-${SLURM_VERSION}.tar.bz2 \
-    && ls -lh /root/rpmbuild/RPMS/x86_64/
+    && ls -lh /root/rpmbuild/RPMS/${RPM_ARCH}/
 
 # ============================================================================
 # Stage 2: Runtime image
@@ -84,6 +92,7 @@ LABEL org.opencontainers.image.source="https://github.com/giovtorres/slurm-docke
       maintainer="Giovanni Torres"
 
 ARG SLURM_VERSION
+ARG TARGETARCH
 
 # Enable CRB and EPEL repositories for runtime dependencies
 RUN set -ex \
@@ -125,17 +134,17 @@ RUN set -ex \
 ARG GOSU_VERSION=1.19
 
 RUN set -ex \
-    && wget -O /usr/local/bin/gosu "https://github.com/tianon/gosu/releases/download/$GOSU_VERSION/gosu-amd64" \
-    && wget -O /usr/local/bin/gosu.asc "https://github.com/tianon/gosu/releases/download/$GOSU_VERSION/gosu-amd64.asc" \
+    && echo "Installing gosu for architecture: ${TARGETARCH}" \
+    && wget -O /usr/local/bin/gosu "https://github.com/tianon/gosu/releases/download/$GOSU_VERSION/gosu-${TARGETARCH}" \
+    && wget -O /usr/local/bin/gosu.asc "https://github.com/tianon/gosu/releases/download/$GOSU_VERSION/gosu-${TARGETARCH}.asc" \
     && export GNUPGHOME="$(mktemp -d)" \
     && gpg --batch --keyserver hkps://keys.openpgp.org --recv-keys B42F6819007F00F88E364FD4036A9C25BF357DD4 \
     && gpg --batch --verify /usr/local/bin/gosu.asc /usr/local/bin/gosu \
     && rm -rf "${GNUPGHOME}" /usr/local/bin/gosu.asc \
     && chmod +x /usr/local/bin/gosu \
     && gosu nobody true
 
-# Copy Slurm RPMs from builder stage
-COPY --from=builder /root/rpmbuild/RPMS/x86_64/*.rpm /tmp/rpms/
+COPY --from=builder /root/rpmbuild/RPMS/*/*.rpm /tmp/rpms/
 
 # Install Slurm RPMs
 RUN set -ex \
@@ -203,7 +212,6 @@ RUN set -ex \
     && rm -rf /tmp/slurm-config
 COPY --chown=slurm:slurm --chmod=0600 examples /root/examples
 
-# Copy entrypoint script
 COPY docker-entrypoint.sh /usr/local/bin/docker-entrypoint.sh
 RUN chmod +x /usr/local/bin/docker-entrypoint.sh
 

README.md

@@ -37,6 +37,17 @@ SLURM_VERSION=24.11.6   # Previous stable release
 
 **Supported versions:** 25.05.x, 24.11.x
 
+## 🏗️ Architecture Support
+
+This project supports both **AMD64 (x86_64)** and **ARM64 (aarch64)**
+architectures. The build system automatically detects your architecture. No
+special configuration is needed - simply build and run:
+
+```bash
+make build
+make up
+```
+
 ## 🚀 Quick Start (Using Make)
 
 The easiest way to get started is using the provided Makefile:
@@ -191,6 +202,23 @@ For more workflows including configuration updates, version switching, and testi
 
 ## ⚙️ Advanced Configuration
 
+### Multi-Architecture Builds
+
+For cross-platform builds or explicit architecture selection (`arm64` or
+`amd64`), use Docker Buildx:
+
+```bash
+docker buildx build \
+  --platform linux/arm64 \
+  --build-arg SLURM_VERSION=25.05.3 \
+  --build-arg TARGETARCH=arm64 \
+  --load \
+  -t slurm-docker-cluster:25.05.3 \
+  .
+```
+
+**Note**: Cross-platform builds use QEMU emulation and may be slower than native builds.
+
 ### Live Configuration Updates
 
 With the `etc_slurm` volume mounted, you can modify configurations without rebuilding:

docker-compose.yml

@@ -25,6 +25,8 @@ services:
       context: .
       args:
         SLURM_VERSION: ${SLURM_VERSION:-25.05.3}
+      cache_from:
+        - slurm-docker-cluster:${SLURM_VERSION:-25.05.3}
     command: ["slurmdbd"]
     container_name: slurmdbd
     hostname: slurmdbd

docker-entrypoint.sh

@@ -81,7 +81,7 @@ then
     if [ -f /proc/self/cgroup ]; then
         # Extract container name from cgroup path
         # Format: 0::/docker/<container_id> or similar
-        CONTAINER_NAME=$(cat /proc/self/cgroup | grep -oP '/docker/\K[^/]+' | head -1)
+        CONTAINER_NAME=$(cat /proc/self/cgroup | sed -n 's|.*/docker/\([^/]*\).*|\1|p' | head -1)
     fi
 
     # If we got a container ID, try to resolve it to a name using the host's /proc

rpmbuild/slurm.rpmmacros

@@ -19,3 +19,6 @@
 
 # MySQL/MariaDB support
 %with_mysql        "--with-mysql_config=/usr/bin"
+
+# Disable InfiniBand/OFED support (not available on ARM64 and not needed for Docker-based testing)
+%_without_ofed 1

run_examples.sh

@@ -38,7 +38,7 @@ case $choice in
             job_name=$(basename "$job_script")
             echo -e "${YELLOW}[SUBMIT]${NC} $job_name"
 
-            job_id=$(docker exec slurmctld bash -c "cd /data && sbatch examples/$job_name 2>&1" | grep -oP 'Submitted batch job \K\d+' || echo "")
+            job_id=$(docker exec slurmctld bash -c "cd /data && sbatch examples/$job_name 2>&1" | sed -n 's/.*Submitted batch job \([0-9][0-9]*\).*/\1/p')
 
             if [ -n "$job_id" ]; then
                 echo -e "${GREEN}  ✓${NC} Job ID: $job_id"
@@ -80,7 +80,7 @@ case $choice in
 
         echo ""
         echo -e "${YELLOW}[SUBMIT]${NC} $job_name"
-        job_id=$(docker exec slurmctld bash -c "cd /data && sbatch examples/$job_name 2>&1" | grep -oP 'Submitted batch job \K\d+' || echo "")
+        job_id=$(docker exec slurmctld bash -c "cd /data && sbatch examples/$job_name 2>&1" | sed -n 's/.*Submitted batch job \([0-9][0-9]*\).*/\1/p')
 
         if [ -n "$job_id" ]; then
             echo -e "${GREEN}  ✓${NC} Job ID: $job_id submitted"

test_cluster.sh

@@ -183,7 +183,7 @@ test_job_submission() {
     print_test "Testing job submission..."
 
     # Submit a simple job
-    JOB_ID=$(docker exec slurmctld bash -c "cd /data && sbatch --wrap='hostname' 2>&1" | grep -oP 'Submitted batch job \K\d+')
+    JOB_ID=$(docker exec slurmctld bash -c "cd /data && sbatch --wrap='hostname' 2>&1" | sed -n 's/.*Submitted batch job \([0-9][0-9]*\).*/\1/p')
 
     if [ -n "$JOB_ID" ]; then
         print_info "  Job ID: $JOB_ID submitted"

update_slurmfiles.sh

@@ -14,7 +14,7 @@ if [ -z "$SLURM_VERSION" ]; then
 fi
 
 # Extract major.minor version (e.g., 25.05 from 25.05.3)
-VERSION_DIR=$(echo "$SLURM_VERSION" | grep -oP '^\d+\.\d+')
+VERSION_DIR=$(echo "$SLURM_VERSION" | cut -d. -f1-2)
 
 restart=false