feat: add network isolation modes to container/run.sh script (#3237)

container/README.md

@@ -167,36 +167,112 @@ The `run.sh` script launches Docker containers with the appropriate configuratio
 - **GPU Management**: Automatic GPU detection and allocation
 - **Volume Mounting**: Workspace and HuggingFace cache mounting
 - **User Management**: Root or user-based container execution
-- **Network Configuration**: Host networking for service communication
+- **Network Configuration**: Configurable networking modes (host, bridge, none, container sharing)
 - **Resource Limits**: Memory, file descriptors, and IPC configuration
 
 **Common Usage Examples:**
 
 ```bash
-# Basic container launch (inference/production)
-./run.sh --image dynamo:latest-vllm
+# Basic container launch (inference/production, runs as root user)
+./run.sh --image dynamo:latest-vllm -v $HOME/.cache:/home/ubuntu/.cache
 
-# Mount workspace for development (use local-dev image for local user permissions)
-./run.sh --image dynamo:latest-vllm-local-dev --mount-workspace
+# Mount workspace for development (use local-dev image for local host user permissions)
+./run.sh --image dynamo:latest-vllm-local-dev --mount-workspace -v $HOME/.cache:/home/ubuntu/.cache
 
 # Use specific image and framework for development
-./run.sh --image v0.1.0.dev.08cc44965-vllm-local-dev --framework vllm --mount-workspace
+./run.sh --image v0.1.0.dev.08cc44965-vllm-local-dev --framework vllm --mount-workspace -v $HOME/.cache:/home/ubuntu/.cache
 
 # Interactive development shell with workspace mounted
-./run.sh --image dynamo:latest-vllm-local-dev --mount-workspace -it -- bash
+./run.sh --image dynamo:latest-vllm-local-dev --mount-workspace -v $HOME/.cache:/home/ubuntu/.cache -it -- bash
 
 # Development with custom environment variables
-./run.sh --image dynamo:latest-vllm-local-dev -e CUDA_VISIBLE_DEVICES=0,1 --mount-workspace
-
-# Production inference without GPU access
-./run.sh --image dynamo:latest-vllm --gpus none
+./run.sh --image dynamo:latest-vllm-local-dev -e CUDA_VISIBLE_DEVICES=0,1 --mount-workspace -v $HOME/.cache:/home/ubuntu/.cache
 
 # Dry run to see docker command
 ./run.sh --dry-run
 
 # Development with custom volume mounts
-./run.sh --image dynamo:latest-vllm-local-dev -v /host/path:/container/path --mount-workspace
+./run.sh --image dynamo:latest-vllm-local-dev -v /host/path:/container/path --mount-workspace -v $HOME/.cache:/home/ubuntu/.cache
+```
+
+### Network Configuration Options
+
+The `run.sh` script supports different networking modes via the `--network` flag (defaults to `host`):
+
+#### Host Networking (Default)
+```bash
+# Same examples with local host user permissions
+./run.sh --image dynamo:latest-vllm-local-dev --network host -v $HOME/.cache:/home/ubuntu/.cache
+./run.sh --image dynamo:latest-vllm-local-dev -v $HOME/.cache:/home/ubuntu/.cache
+```
+**Use cases:**
+- High-performance ML inference (default for GPU workloads)
+- Services that need direct host port access
+- Maximum network performance with minimal overhead
+- Sharing services with the host machine (NATS, etcd, etc.)
+
+**⚠️ Port Sharing Limitation:** Host networking shares all ports with the host machine, which means you can only run **one instance** of services like NATS (port 4222) or etcd (port 2379) across all containers and the host.
+
+#### Bridge Networking (Isolated)
+```bash
+# CI/testing with isolated bridge networking and host cache sharing
+./run.sh --image dynamo:latest-vllm --mount-workspace --network bridge -v $HOME/.cache:/home/ubuntu/.cache
+```
+**Use cases:**
+- Secure isolation from host network
+- CI/CD pipelines requiring complete isolation
+- When you need absolute control of ports
+- Exposing specific services to host while maintaining isolation
+
+**Note:** For port sharing with the host, use the `--port` or `-p` option with format `host_port:container_port` (e.g., `--port 8000:8000` or `-p 9081:8081`) to expose specific container ports to the host.
+
+#### No Networking ⚠️ **LIMITED FUNCTIONALITY**
+```bash
+# Complete network isolation - no external connectivity
+./run.sh --image dynamo:latest-vllm --network none --mount-workspace -v $HOME/.cache:/home/ubuntu/.cache
+
+# Same with local user permissions
+./run.sh --image dynamo:latest-vllm-local-dev --network none --mount-workspace -v $HOME/.cache:/home/ubuntu/.cache
 ```
+**⚠️ WARNING: `--network none` severely limits Dynamo functionality:**
+- **No model downloads** - HuggingFace models cannot be downloaded
+- **No API access** - Cannot reach external APIs or services
+- **No distributed inference** - Multi-node setups won't work
+- **No monitoring/logging** - External monitoring systems unreachable
+- **Limited debugging** - Cannot access external debugging tools
+
+**Very limited use cases:**
+- Pre-downloaded models with purely local processing
+- Air-gapped security environments (models must be pre-staged)
+
+#### Container Network Sharing
+Use `--network container:name` to share the network namespace with another container.
+
+**Use cases:**
+- Sidecar patterns (logging, monitoring, caching)
+- Service mesh architectures
+- Sharing network namespaces between related containers
+
+See Docker documentation for `--network container:name` usage.
+
+#### Custom Networks
+Use custom Docker networks for multi-container applications. Create with `docker network create` and specify with `--network network-name`.
+
+**Use cases:**
+- Multi-container applications
+- Service discovery by container name
+
+See Docker documentation for custom network creation and management.
+
+#### Network Mode Comparison
+
+| Mode | Performance | Security | Use Case | Dynamo Compatibility | Port Sharing | Port Publishing |
+|------|-------------|----------|----------|---------------------|---------------|-----------------|
+| `host` | Highest | Lower | ML/GPU workloads, high-performance services | ✅ Full | ⚠️ **Shared with host** (one NATS/etcd only) | ❌ Not needed |
+| `bridge` | Good | Higher | General web services, controlled port exposure | ✅ Full | ✅ Isolated ports | ✅ `-p host:container` |
+| `none` | N/A | Highest | Air-gapped environments only | ⚠️ **Very Limited** | ✅ No network | ❌ No network |
+| `container:name` | Good | Medium | Sidecar patterns, shared network stacks | ✅ Full | ⚠️ Shared with target container | ❌ Use target's ports |
+| Custom networks | Good | Medium | Multi-container applications | ✅ Full | ✅ Isolated ports | ✅ `-p host:container` |
 
 ## Workflow Examples
 
@@ -206,30 +282,51 @@ The `run.sh` script launches Docker containers with the appropriate configuratio
 ./build.sh --framework vllm --target local-dev
 
 # 2. Run development container using the local-dev image
-./run.sh --image dynamo:latest-vllm-local-dev --mount-workspace -it
+./run.sh --image dynamo:latest-vllm-local-dev --mount-workspace -v $HOME/.cache:/home/ubuntu/.cache -it
 
 # 3. Inside container, run inference (requires both frontend and backend)
 # Start frontend
 python -m dynamo.frontend &
 
 # Start backend (vLLM example)
-python -m dynamo.vllm --model Qwen/Qwen3-0.6B --gpu-memory-utilization 0.50 &
+python -m dynamo.vllm --model Qwen/Qwen3-0.6B --gpu-memory-utilization 0.20 &
 ```
 
 ### Production Workflow
 ```bash
 # 1. Build production image
 ./build.sh --framework vllm --release-build
 
-# 2. Run production container
-./run.sh --image dynamo:latest-vllm-local-dev --gpus all
+# 2. Run production container (runs as root)
+./run.sh --image dynamo:latest-vllm --gpus all
 ```
 
-### Testing Workflow
+### CI/CD Workflow
 ```bash
-# 1. Build with no cache for clean build
+# 1. Build image for CI
 ./build.sh --framework vllm --no-cache
 
-# 2. Test container functionality (--image defaults to dynamo:latest-vllm)
-./run.sh --mount-workspace -it -- python -m pytest tests/
+# 2. Run tests with network isolation for reproducible results
+./run.sh --image dynamo:latest-vllm --mount-workspace --network bridge -v $HOME/.cache:/home/ubuntu/.cache -- python -m pytest tests/
+
+# 3. Inside the container with bridge networking, start services
+# Note: Services are only accessible from the same container - no port conflicts with host
+nats-server -js &
+etcd --listen-client-urls http://0.0.0.0:2379 --advertise-client-urls http://0.0.0.0:2379 --data-dir /tmp/etcd &
+python -m dynamo.frontend &
+
+# 4. Start worker backend (choose one framework):
+# vLLM
+DYN_SYSTEM_ENABLED=true DYN_SYSTEM_PORT=8081 python -m dynamo.vllm --model Qwen/Qwen3-0.6B --gpu-memory-utilization 0.20 --enforce-eager --no-enable-prefix-caching --max-num-seqs 64 &
+
+# SGLang
+DYN_SYSTEM_ENABLED=true DYN_SYSTEM_PORT=8081 python -m dynamo.sglang --model Qwen/Qwen3-0.6B --mem-fraction-static 0.20 --max-running-requests 64 &
+
+# TensorRT-LLM
+DYN_SYSTEM_ENABLED=true DYN_SYSTEM_PORT=8081 python -m dynamo.trtllm --model Qwen/Qwen3-0.6B --free-gpu-memory-fraction 0.20 --max-num-tokens 8192 --max-batch-size 64 &
 ```
+
+**Framework-Specific GPU Memory Arguments:**
+- **vLLM**: `--gpu-memory-utilization 0.20` (use 20% GPU memory), `--enforce-eager` (disable CUDA graphs), `--no-enable-prefix-caching` (save memory), `--max-num-seqs 64` (max concurrent sequences)
+- **SGLang**: `--mem-fraction-static 0.20` (20% GPU memory for static allocation), `--max-running-requests 64` (max concurrent requests)
+- **TensorRT-LLM**: `--free-gpu-memory-fraction 0.20` (reserve 20% GPU memory), `--max-num-tokens 8192` (max tokens in batch), `--max-batch-size 64` (max batch size)

container/run.sh

@@ -36,13 +36,15 @@ DEFAULT_HF_CACHE=${SOURCE_DIR}/.cache/huggingface
 GPUS="all"
 PRIVILEGED=
 VOLUME_MOUNTS=
+PORT_MAPPINGS=
 MOUNT_WORKSPACE=
 ENVIRONMENT_VARIABLES=
 REMAINING_ARGS=
 INTERACTIVE=
 USE_NIXL_GDS=
 RUNTIME=nvidia
 WORKDIR=/workspace
+NETWORK=host
 
 get_options() {
     while :; do
@@ -148,6 +150,14 @@ get_options() {
                 missing_requirement "$1"
             fi
             ;;
+        -p|--port)
+            if [ "$2" ]; then
+                PORT_MAPPINGS+=" -p $2 "
+                shift
+            else
+                missing_requirement "$1"
+            fi
+            ;;
         -e)
             if [ "$2" ]; then
                 ENVIRONMENT_VARIABLES+=" -e $2 "
@@ -165,6 +175,14 @@ get_options() {
         --use-nixl-gds)
             USE_NIXL_GDS=TRUE
             ;;
+        --network)
+            if [ "$2" ]; then
+                NETWORK=$2
+                shift
+            else
+                missing_requirement "$1"
+            fi
+            ;;
         --dry-run)
             RUN_PREFIX="echo"
             echo ""
@@ -304,7 +322,12 @@ show_help() {
     echo "  [--hf-cache directory to volume mount as the hf cache, default is NONE unless mounting workspace]"
     echo "  [--gpus gpus to enable, default is 'all', 'none' disables gpu support]"
     echo "  [--use-nixl-gds add volume mounts and capabilities needed for NVIDIA GPUDirect Storage]"
+    echo "  [--network network mode for container, default is 'host']"
+    echo "           Options: 'host' (default), 'bridge', 'none', 'container:name'"
+    echo "           Examples: --network bridge (isolated), --network none (no network - WARNING: breaks most functionality)"
+    echo "                    --network container:redis (share network with 'redis' container)"
     echo "  [-v add volume mount]"
+    echo "  [-p|--port add port mapping (host_port:container_port)]"
     echo "  [-e add environment variable]"
     echo "  [--mount-workspace set up for local development]"
     echo "  [-- stop processing and pass remaining args as command to docker run]"
@@ -335,14 +358,15 @@ ${RUN_PREFIX} docker run \
     ${GPU_STRING} \
     ${INTERACTIVE} \
     ${RM_STRING} \
-    --network host \
+    --network "$NETWORK" \
     ${RUNTIME:+--runtime "$RUNTIME"} \
     --shm-size=10G \
     --ulimit memlock=-1 \
     --ulimit stack=67108864 \
     --ulimit nofile=65536:65536 \
     ${ENVIRONMENT_VARIABLES} \
     ${VOLUME_MOUNTS} \
+    ${PORT_MAPPINGS} \
     -w "$WORKDIR" \
     --cap-add CAP_SYS_PTRACE \
     ${NIXL_GDS_CAPS} \