feat(docker): autoheal

Dockerfile

@@ -26,7 +26,7 @@ RUN cargo build --release
 FROM debian:bookworm
 # Install runtime dependencies
 RUN apt-get -y update && \
-    apt-get install -y openssl ca-certificates tini &&\
+    apt-get install -y openssl ca-certificates tini curl &&\
     apt-get autoremove -y; \
     apt-get clean; \
     rm -rf /var/lib/apt/lists/*

Makefile

@@ -1,5 +1,5 @@
 # ============================================================================ #
-#                         STARKNET NODE BENCHMAR RUNNER                        #
+#                              STARKNET NODE RUNNER                            #
 # ============================================================================ #
 
 define HELP
@@ -66,21 +66,27 @@ Targets:
 endef
 export HELP
 
-SECRETS  := .secrets/rpc_api.secret
+SECRETS        := .secrets/rpc_api.secret
+DB_PATH        := /var/lib/madara
+
+DOCKER_COMPOSE := docker compose -f compose.yaml
+DOCKER_TAG     := madara:latest
+DOCKER_IMAGE   := ghcr.io/madara-alliance/$(DOCKER_TAG)
+DOCKER_GZ      := image.tar.gz
 
 # dim white italic
-DIM      := \033[2;3;37m
+DIM            := \033[2;3;37m
 
 # bold cyan
-INFO     := \033[1;36m
+INFO           := \033[1;36m
 
 # bold green
-PASS     := \033[1;32m
+PASS           := \033[1;32m
 
 # bold red
-WARN     := \033[1;31m
+WARN           := \033[1;31m
 
-RESET    := \033[0m
+RESET          := \033[0m
 
 .PHONY: all
 all: help
@@ -92,27 +98,27 @@ help:
 .PHONY: start
 start: images $(SECRETS)
 	@echo -e "$(DIM)running$(RESET) $(PASS)madara$(RESET)"
-	@docker compose -f compose.yaml up -d
+	@$(DOCKER_COMPOSE) up -d
 
 .PHONY: stop
 stop:
 	@echo -e "$(DIM)stopping$(RESET) $(WARN)madara$(RESET)"
-	@docker compose -f compose.yaml stop
+	@$(DOCKER_COMPOSE) stop
 
 .PHONY: logs
 logs:
 	@echo -e "$(DIM)logs for$(RESET) $(INFO)madara$(RESET)";
-	@docker compose -f compose.yaml logs -f -n 100;
+	@$(DOCKER_COMPOSE) logs -f -n 100 madara;
 
 .PHONY: images
-images: image.tar.gz
+images: $(DOCKER_GZ)
 
-image.tar.gz:
+$(DOCKER_GZ):
 	@echo -e "$(DIM)downloading$(RESET) $(PASS)madara$(RESET)"
-	@docker pull ghcr.io/madara-alliance/madara:latest
-	@docker tag ghcr.io/madara-alliance/madara:latest madara:latest
-	@docker rmi ghcr.io/madara-alliance/madara:latest
-	@docker image save -o image.tar.gz madara:latest
+	@docker pull $(DOCKER_IMAGE)
+	@docker tag $(DOCKER_IMAGE) $(DOCKER_TAG)
+	@docker rmi $(DOCKER_IMAGE)
+	@docker image save -o $(DOCKER_GZ) $(DOCKER_TAG)
 
 .PHONY: clean
 clean: stop
@@ -133,12 +139,12 @@ clean-db:
 	esac
 	@make --silent clean
 	@echo -e "$(DIM)removing madara database on host$(RESET)"
-	@sudo rm -rf /var/lib/madara;
+	@sudo rm -rf $(DB_PATH);
 
 .PHONY: fclean
 fclean: clean-db
 	@echo -e "$(DIM)removing local images tar.gz$(RESET)"
-	@rm -rf image.tar.gz
+	@rm -rf $(DOCKER_GZ)
 	@echo -e "$(WARN)artefacts cleaned$(RESET)"
 
 .PHONY: restart

README.md

@@ -55,6 +55,7 @@ Ensure you have all the necessary dependencies available on your host system.
 | ---------- | ---------- | ----------------------------------------------------------------- |
 | Rust       | rustc 1.81 | `curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs \| sh` |
 | Clang      | Latest     | `sudo apt-get install clang`                                      |
+| Openssl    | 0.10       | `sudo apt install openssl`                                        |
 
 Once all dependencies are satisfied, you can clone the Madara repository:
 
@@ -128,7 +129,7 @@ cargo run --release --        \
 ```
 
 > [!NOTE]
-> Head to the [Configuration](#configuration) section to learn more about
+> Head to the [Configuration](#%EF%B8%8F-configuration) section to learn more about
 > customizing your node.
 
 #### 4. Presets
@@ -225,6 +226,14 @@ make clean-db # Removes the madara db, including files on the host
 make restart  # Restarts the madara node
 ```
 
+> [!IMPORTANT]
+> By default, `make start` will try and restart Madara indefinitely if it is
+> found to be unhealthy using [docker autoheal](https://github.com/willfarrell/docker-autoheal).
+> This is done by checking the availability of `http://localhost:9944/health`,
+> which means your container will be marked as `unhealthy` and restart if you
+> have disabled the RPC service! You can run `watch docker ps` to monitor the
+> health of your containers.
+
 To change runtime arguments, you can update the script in `madara-runner.sh`:
 
 ```bash
@@ -349,9 +358,8 @@ Here is a list of all the supported methods with their current status:
 | ✅     | `starknet_getEvents`                       |
 | ✅     | `starknet_getNonce`                        |
 | ✅     | `starknet_getCompiledCasm` (v0.8.0)        |
-| 🚧     | `getMessageStatus` (v0.8.0)                |
-| 🚧     | `getStorageProof` (v0.8.0)                 |
-| ❌     | `getTransactionStatus` (v0.8.0)            |
+| 🚧     | `starknet_getMessageStatus` (v0.8.0)       |
+| 🚧     | `starknet_getStorageProof` (v0.8.0)        |
 
 </details>
 

compose.yaml

@@ -7,21 +7,33 @@ services:
     cpu_count: 4
     mem_limit: "16gb"
     ports:
-      - "9944"
+      - 9944:9944
+    labels:
+      - "autoheal=true"
     environment:
-      RPC_API_KEY_FILE: /run/secrets/rpc_api_key
+      - RPC_API_KEY_FILE=/run/secrets/rpc_api_key
     secrets:
       - rpc_api_key
     volumes:
       - /var/lib/madara:/tmp/madara
-      - type: bind
-        source: ./madara-runner.sh
-        target: /usr/local/bin/runner.sh
-        read_only: true
-        bind:
-          create_host_path: false
+      - ./madara-runner.sh:/usr/local/bin/runner.sh:ro
     entrypoint:
       - /usr/local/bin/runner.sh
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:9944/health"]
+      interval: 10s
+      timeout: 5s
+      retries: 3
+      start_period: 10s
+    restart: on-failure:3
+
+  autoheal:
+    image: willfarrell/autoheal:latest
+    tty: true
+    restart: always
+    volumes:
+      - /etc/localtime:/etc/localtime:ro
+      - /var/run/docker.sock:/var/run/docker.sock
 
 secrets:
   rpc_api_key:

crates/tests/src/rpc/read.rs

@@ -41,12 +41,12 @@ mod test_rpc_read_calls {
     }
     impl Drop for SharedMadaraInstance {
         fn drop(&mut self) {
-            let mut guard = MADARA_HANDLE_COUNT.lock().expect("poisoned lock");
-            *guard -= 1;
-            if *guard == 0 {
-                // :/
-                tokio::task::spawn_blocking(|| *MADARA.blocking_lock() = None);
-            }
+            // let mut guard = MADARA_HANDLE_COUNT.lock().expect("poisoned lock");
+            // *guard -= 1;
+            // if *guard == 0 {
+            //     // :/
+            //     tokio::task::spawn_blocking(|| *MADARA.blocking_lock() = None);
+            // }
         }
     }
     #[allow(clippy::await_holding_lock)]
@@ -650,6 +650,10 @@ mod test_rpc_read_calls {
         };
         let expected_txn_status = TransactionStatus::AcceptedOnL2(TransactionExecutionStatus::Succeeded);
 
+        // TODO: The shared madara state needs a rework as we only run these
+        // tests with `--test-threads=1`. These tests
+        tokio::task::spawn_blocking(|| *MADARA.blocking_lock() = None);
+
         assert_eq!(txn_status, expected_txn_status);
     }
 

madara-runner.sh

@@ -1,10 +1,14 @@
 #!/bin/sh
-export RPC_API_KEY=$(cat $RPC_API_KEY_FILE)
+if [ -f "$RPC_API_KEY_FILE" ]; then
+  export RPC_API_KEY=$(cat "$RPC_API_KEY_FILE")
+else
+  echo "Error: RPC_API_KEY_FILE not found!" >&2
+  exit 1
+fi
 
-./madara                   \
+tini -- ./madara           \
 	--name madara            \
 	--network mainnet        \
 	--rpc-external           \
-	--rpc-cors all           \
 	--full                   \
 	--l1-endpoint $RPC_API_KEY