Dboreham/deploy volume control (#401)

* Implement volume control

* Deploy test

* Add test for volumes

* Enable CI for deploy test

.gitea/workflows/test-deploy.yml

@@ -0,0 +1,39 @@
+name: Deploy Test
+
+on:
+  pull_request:
+    branches: '*'
+  push:
+    branches:
+      - main
+      - ci-test
+
+# Needed until we can incorporate docker startup into the executor container
+env:
+  DOCKER_HOST: unix:///var/run/dind.sock
+
+jobs:
+  test:
+    name: "Run deploy test suite"
+    runs-on: ubuntu-latest
+    steps:
+      - name: "Clone project repository"
+        uses: actions/checkout@v3
+      - name: "Install Python"
+        uses: cerc-io/setup-python@v4
+        with:
+          python-version: '3.8'
+      - name: "Print Python version"
+        run: python3 --version
+      - name: "Install shiv"
+        run: pip install shiv
+      - name: "Generate build version file"
+        run: ./scripts/create_build_tag_file.sh
+      - name: "Build local shiv package"
+        run: ./scripts/build_shiv_package.sh
+      - name: Start dockerd # Also needed until we can incorporate into the executor
+        run: |
+          dockerd -H $DOCKER_HOST --userland-proxy=false &
+          sleep 5
+      - name: "Run smoke tests"
+        run: ./tests/deploy-test/run-smoke-test.sh

.github/workflows/test-deploy.yml

@@ -0,0 +1,29 @@
+name: Deploy Test
+
+on:
+  pull_request:
+    branches: '*'
+  push:
+    branches: '*'
+
+jobs:
+  test:
+    name: "Run deploy test suite"
+    runs-on: ubuntu-latest
+    steps:
+      - name: "Clone project repository"
+        uses: actions/checkout@v3
+      - name: "Install Python"
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.8'
+      - name: "Print Python version"
+        run: python3 --version
+      - name: "Install shiv"
+        run: pip install shiv
+      - name: "Generate build version file"
+        run: ./scripts/create_build_tag_file.sh
+      - name: "Build local shiv package"
+        run: ./scripts/build_shiv_package.sh
+      - name: "Run smoke tests"
+        run: ./tests/deploy-test/run-smoke-test.sh

app/data/compose/docker-compose-fixturenet-eth.yml

@@ -2,6 +2,7 @@ version: '3.7'
 
 services:
   fixturenet-eth-bootnode-geth:
+    restart: always
     hostname: fixturenet-eth-bootnode-geth
     env_file:
       - ../config/fixturenet-eth/fixturenet-eth.env
@@ -15,6 +16,7 @@ services:
       - "30303"
 
   fixturenet-eth-geth-1:
+    restart: always
     hostname: fixturenet-eth-geth-1
     cap_add:
       - SYS_PTRACE
@@ -42,6 +44,7 @@ services:
       - "6060"
 
   fixturenet-eth-geth-2:
+    restart: always
     hostname: fixturenet-eth-geth-2
     healthcheck:
       test: ["CMD", "nc", "-v", "localhost", "8545"]
@@ -60,12 +63,14 @@ services:
       - fixturenet_eth_geth_2_data:/root/ethdata
 
   fixturenet-eth-bootnode-lighthouse:
+    restart: always
     hostname: fixturenet-eth-bootnode-lighthouse
     environment:
       RUN_BOOTNODE: "true"
     image: cerc/fixturenet-eth-lighthouse:local
 
   fixturenet-eth-lighthouse-1:
+    restart: always
     hostname: fixturenet-eth-lighthouse-1
     healthcheck:
       test: ["CMD", "wget", "--tries=1", "--connect-timeout=1", "--quiet", "-O", "-", "http://localhost:8001/eth/v2/beacon/blocks/head"]
@@ -91,6 +96,7 @@ services:
       - "8001"
 
   fixturenet-eth-lighthouse-2:
+    restart: always
     hostname: fixturenet-eth-lighthouse-2
     healthcheck:
       test: ["CMD", "wget", "--tries=1", "--connect-timeout=1", "--quiet", "-O", "-", "http://localhost:8001/eth/v2/beacon/blocks/head"]

app/data/compose/docker-compose-fixturenet-laconicd.yml

@@ -1,10 +1,11 @@
-version: "3.2"
 services:
   laconicd:
     restart: unless-stopped
     image: cerc/laconicd:local
     command: ["sh", "/docker-entrypoint-scripts.d/create-fixturenet.sh"]
     volumes:
+      # The cosmos-sdk node's database directory:
+      - laconicd-data:/root/.laconicd/data
       # TODO: look at folding these scripts into the container
       - ../config/fixturenet-laconicd/create-fixturenet.sh:/docker-entrypoint-scripts.d/create-fixturenet.sh
       - ../config/fixturenet-laconicd/export-mykey.sh:/docker-entrypoint-scripts.d/export-mykey.sh

app/data/compose/docker-compose-fixturenet-optimism.yml

@@ -5,6 +5,7 @@ services:
   # Creates / updates the configuration for L1 contracts deployment
   # Deploys the L1 smart contracts (outputs to volume l1_deployment)
   fixturenet-optimism-contracts:
+    restart: on-failure
     hostname: fixturenet-optimism-contracts
     image: cerc/optimism-contracts:local
     env_file:
@@ -35,6 +36,7 @@ services:
 
   # Generates the config files required for L2 (outputs to volume l2_config)
   op-node-l2-config-gen:
+    restart: on-failure
     image: cerc/optimism-op-node:local
     depends_on:
       fixturenet-optimism-contracts:
@@ -54,6 +56,7 @@ services:
 
   # Initializes and runs the L2 execution client  (outputs to volume l2_geth_data)
   op-geth:
+    restart: always
     image: cerc/optimism-l2geth:local
     depends_on:
       op-node-l2-config-gen:
@@ -76,6 +79,7 @@ services:
 
   # Runs the L2 consensus client (Sequencer node)
   op-node:
+    restart: always
     image: cerc/optimism-op-node:local
     depends_on:
       op-geth:
@@ -103,6 +107,7 @@ services:
 
   # Runs the batcher (takes transactions from the Sequencer and publishes them to L1)
   op-batcher:
+    restart: always
     image: cerc/optimism-op-batcher:local
     depends_on:
       op-node:
@@ -129,6 +134,7 @@ services:
 
   # Runs the proposer (periodically submits new state roots to L1)
   op-proposer:
+    restart: always
     image: cerc/optimism-op-proposer:local
     depends_on:
       op-node:

app/data/compose/docker-compose-foundry.yml

@@ -1,6 +1,7 @@
 # Add-on pod to include foundry tooling within a fixturenet
 services:
   foundry:
+    restart: always
     image: cerc/foundry:local
     command: ["while :; do sleep 600; done"]
     volumes:

app/data/compose/docker-compose-test.yml

@@ -1,7 +1,13 @@
-version: "3.2"
 services:
   test:
     image: cerc/test-container:local
     restart: always
+    environment:
+      CERC_SCRIPT_DEBUG: ${CERC_SCRIPT_DEBUG}
+    volumes:
+      - test-data:/var
     ports:
       - "80"
+
+volumes:
+  test-data:

app/data/container-build/cerc-fixturenet-eth-geth/Dockerfile

@@ -6,7 +6,7 @@ RUN go install github.com/go-delve/delve/cmd/dlv@latest
 
 FROM cerc/go-ethereum:local as geth
 
-FROM alpine:latest
+FROM alpine:3.17
 RUN apk add --no-cache python3 python3-dev py3-pip curl wget jq build-base gettext libintl openssl bash bind-tools postgresql-client
 
 COPY --from=delve  /go/bin/dlv /usr/local/bin/

app/data/container-build/cerc-fixturenet-eth-lighthouse/genesis/cl/bootnode.sh

@@ -13,22 +13,26 @@ DEBUG_LEVEL=${1:-info}
 
 echo "Starting bootnode"
 
-if [ ! -f "$DATADIR/bootnode/enr.dat" ]; then
-  echo "Generating bootnode enr"
-  lcli \
-    generate-bootnode-enr \
-    --ip $ENR_IP \
-    --udp-port $BOOTNODE_PORT \
-    --tcp-port $BOOTNODE_PORT \
-    --genesis-fork-version $GENESIS_FORK_VERSION \
-    --output-dir $DATADIR/bootnode
-
-  bootnode_enr=`cat $DATADIR/bootnode/enr.dat`
-  echo "- $bootnode_enr" > $TESTNET_DIR/boot_enr.yaml
-
-  echo "Generated bootnode enr and written to $TESTNET_DIR/boot_enr.yaml"
+# Clean up existing ENR dir to avoid node connectivity issues on a restart
+if [ -d "$DATADIR/bootnode" ]; then
+  echo "Removing existing bootnode enr directory"
+  rm -r "$DATADIR/bootnode"
 fi
 
+echo "Generating bootnode enr"
+lcli \
+  generate-bootnode-enr \
+  --ip $ENR_IP \
+  --udp-port $BOOTNODE_PORT \
+  --tcp-port $BOOTNODE_PORT \
+  --genesis-fork-version $GENESIS_FORK_VERSION \
+  --output-dir $DATADIR/bootnode
+
+bootnode_enr=`cat $DATADIR/bootnode/enr.dat`
+echo "- $bootnode_enr" > $TESTNET_DIR/boot_enr.yaml
+
+echo "Generated bootnode enr and written to $TESTNET_DIR/boot_enr.yaml"
+
 exec lighthouse boot_node \
     --testnet-dir $TESTNET_DIR \
     --port $BOOTNODE_PORT \

app/data/container-build/cerc-test-container/run.sh

@@ -1,11 +1,14 @@
-#!/bin/sh
-
+#!/usr/bin/env bash
+set -e
+if [ -n "$CERC_SCRIPT_DEBUG" ]; then
+  set -x
+fi
 # Test if the container's filesystem is old (run previously) or new
 EXISTSFILENAME=/var/exists
 echo "Test container starting"
 if [[ -f "$EXISTSFILENAME" ]];
 then
-    TIMESTAMP = `cat $EXISTSFILENAME`
+    TIMESTAMP=`cat $EXISTSFILENAME`
     echo "Filesystem is old, created: $TIMESTAMP" 
 else
     echo "Filesystem is fresh"

app/deploy_system.py

@@ -28,108 +28,146 @@ import importlib.resources
 from pathlib import Path
 from .util import include_exclude_check, get_parsed_stack_config
 
+class DeployCommandContext(object):
+    def __init__(self, cluster_context, docker):
+        self.cluster_context = cluster_context
+        self.docker = docker
 
-@click.command()
+
+@click.group()
 @click.option("--include", help="only start these components")
 @click.option("--exclude", help="don\'t start these components")
 @click.option("--env-file", help="env file to be used")
 @click.option("--cluster", help="specify a non-default cluster name")
-@click.argument('command', required=True)  # help: command: up|down|ps
-@click.argument('extra_args', nargs=-1)  # help: command: up|down|ps <service1> <service2>
 @click.pass_context
-def command(ctx, include, exclude, env_file, cluster, command, extra_args):
+def command(ctx, include, exclude, env_file, cluster):
     '''deploy a stack'''
 
-    # TODO: implement option exclusion and command value constraint lost with the move from argparse to click
-
-    debug = ctx.obj.debug
-    quiet = ctx.obj.quiet
-    verbose = ctx.obj.verbose
-    local_stack = ctx.obj.local_stack
-    dry_run = ctx.obj.dry_run
-    stack = ctx.obj.stack
-
-    cluster_context = _make_cluster_context(ctx.obj, include, exclude, cluster)
+    cluster_context = _make_cluster_context(ctx.obj, include, exclude, cluster, env_file)
 
     # See: https://gabrieldemarmiesse.github.io/python-on-whales/sub-commands/compose/
-    docker = DockerClient(compose_files=cluster_context.compose_files, compose_project_name=cluster_context.cluster, compose_env_file=env_file)
+    docker = DockerClient(compose_files=cluster_context.compose_files, compose_project_name=cluster_context.cluster,
+                          compose_env_file=cluster_context.env_file)
 
+    ctx.obj = DeployCommandContext(cluster_context, docker)
+    # Subcommand is executed now, by the magic of click
+
+
+@command.command()
+@click.argument('extra_args', nargs=-1)  # help: command: up <service1> <service2>
+@click.pass_context
+def up(ctx, extra_args):
+    global_context = ctx.parent.parent.obj
     extra_args_list = list(extra_args) or None
+    if not global_context.dry_run:
+        cluster_context = ctx.obj.cluster_context
+        container_exec_env = _make_runtime_env(global_context)
+        for attr, value in container_exec_env.items():
+            os.environ[attr] = value
+        if global_context.verbose:
+            print(f"Running compose up with container_exec_env: {container_exec_env}, extra_args: {extra_args_list}")
+        for pre_start_command in cluster_context.pre_start_commands:
+            _run_command(ctx.obj, cluster_context.cluster, pre_start_command)
+        ctx.obj.docker.compose.up(detach=True, services=extra_args_list)
+        for post_start_command in cluster_context.post_start_commands:
+            _run_command(ctx.obj, cluster_context.cluster, post_start_command)
+        _orchestrate_cluster_config(ctx.obj, cluster_context.config, ctx.obj.docker, container_exec_env)
 
-    if not dry_run:
-        if command == "up":
-            container_exec_env = _make_runtime_env(ctx.obj)
-            for attr, value in container_exec_env.items():
-                os.environ[attr] = value
-            if verbose:
-                print(f"Running compose up with container_exec_env: {container_exec_env}, extra_args: {extra_args_list}")
-            for pre_start_command in cluster_context.pre_start_commands:
-                _run_command(ctx.obj, cluster_context.cluster, pre_start_command)
-            docker.compose.up(detach=True, services=extra_args_list)
-            for post_start_command in cluster_context.post_start_commands:
-                _run_command(ctx.obj, cluster_context.cluster, post_start_command)
 
-            _orchestrate_cluster_config(ctx.obj, cluster_context.config, docker, container_exec_env)
+@command.command()
+@click.option("--delete-volumes/--preserve-volumes", default=False, help="delete data volumes")
+@click.argument('extra_args', nargs=-1)  # help: command: down<service1> <service2>
+@click.pass_context
+def down(ctx, delete_volumes, extra_args):
+    global_context = ctx.parent.parent.obj
+    extra_args_list = list(extra_args) or None
+    if not global_context.dry_run:
+        if global_context.verbose:
+            print("Running compose down")
+        timeout_arg = None
+        if extra_args_list:
+            timeout_arg = extra_args_list[0]
+        # Specify shutdown timeout (default 10s) to give services enough time to shutdown gracefully
+        ctx.obj.docker.compose.down(timeout=timeout_arg, volumes=delete_volumes)
 
-        elif command == "down":
-            if verbose:
-                print("Running compose down")
 
-            timeout_arg = None
-            if extra_args_list:
-                timeout_arg=extra_args_list[0]
+@command.command()
+@click.pass_context
+def ps(ctx):
+    global_context = ctx.parent.parent.obj
+    if not global_context.dry_run:
+        if global_context.verbose:
+            print("Running compose ps")
+        container_list = ctx.obj.docker.compose.ps()
+        if len(container_list) > 0:
+            print("Running containers:")
+            for container in container_list:
+                print(f"id: {container.id}, name: {container.name}, ports: ", end="")
+                ports = container.network_settings.ports
+                comma = ""
+                for port_mapping in ports.keys():
+                    mapping = ports[port_mapping]
+                    print(comma, end="")
+                    if mapping is None:
+                        print(f"{port_mapping}", end="")
+                    else:
+                        print(f"{mapping[0]['HostIp']}:{mapping[0]['HostPort']}->{port_mapping}", end="")
+                    comma = ", "
+                print()
+        else:
+            print("No containers running")
 
-            # Specify shutdown timeout (default 10s) to give services enough time to shutdown gracefully
-            docker.compose.down(timeout=timeout_arg)
-        elif command == "exec":
-            if extra_args_list is None or len(extra_args_list) < 2:
-                print("Usage: exec <service> <cmd>")
-                sys.exit(1)
-            service_name = extra_args_list[0]
-            command_to_exec = ["sh", "-c"] + extra_args_list[1:]
-            container_exec_env = _make_runtime_env(ctx.obj)
-            if verbose:
-                print(f"Running compose exec {service_name} {command_to_exec}")
-            try:
-                docker.compose.execute(service_name, command_to_exec, envs=container_exec_env)
-            except DockerException as error:
-                print(f"container command returned error exit status")
-        elif command == "port":
-            if extra_args_list is None or len(extra_args_list) < 2:
-                print("Usage: port <service> <exposed-port>")
-                sys.exit(1)
-            service_name = extra_args_list[0]
-            exposed_port = extra_args_list[1]
-            if verbose:
-                print(f"Running compose port {service_name} {exposed_port}")
-            mapped_port_data = docker.compose.port(service_name, exposed_port)
-            print(f"{mapped_port_data[0]}:{mapped_port_data[1]}")
-        elif command == "ps":
-            if verbose:
-                print("Running compose ps")
-            container_list = docker.compose.ps()
-            if len(container_list) > 0:
-                print("Running containers:")
-                for container in container_list:
-                    print(f"id: {container.id}, name: {container.name}, ports: ", end="")
-                    ports = container.network_settings.ports
-                    comma = ""
-                    for port_mapping in ports.keys():
-                        mapping = ports[port_mapping]
-                        print(comma, end="")
-                        if mapping is None:
-                            print(f"{port_mapping}", end="")
-                        else:
-                            print(f"{mapping[0]['HostIp']}:{mapping[0]['HostPort']}->{port_mapping}", end="")
-                        comma = ", "
-                    print()
-            else:
-                print("No containers running")
-        elif command == "logs":
-            if verbose:
-                print("Running compose logs")
-            logs_output = docker.compose.logs(services=extra_args_list if extra_args_list is not None else [])
-            print(logs_output)
+
+@command.command()
+@click.argument('extra_args', nargs=-1)  # help: command: port <service1> <service2>
+@click.pass_context
+def port(ctx, extra_args):
+    global_context = ctx.parent.parent.obj
+    extra_args_list = list(extra_args) or None
+    if not global_context.dry_run:
+        if extra_args_list is None or len(extra_args_list) < 2:
+            print("Usage: port <service> <exposed-port>")
+            sys.exit(1)
+        service_name = extra_args_list[0]
+        exposed_port = extra_args_list[1]
+        if ctx.parent.obj.verbose:
+            print(f"Running compose port {service_name} {exposed_port}")
+        mapped_port_data = ctx.obj.docker.compose.port(service_name, exposed_port)
+        print(f"{mapped_port_data[0]}:{mapped_port_data[1]}")
+
+
+@command.command()
+@click.argument('extra_args', nargs=-1)  # help: command: exec <service> <command>
+@click.pass_context
+def exec(ctx, extra_args):
+    global_context = ctx.parent.parent.obj
+    extra_args_list = list(extra_args) or None
+    if not global_context.dry_run:
+        if extra_args_list is None or len(extra_args_list) < 2:
+            print("Usage: exec <service> <cmd>")
+            sys.exit(1)
+        service_name = extra_args_list[0]
+        command_to_exec = ["sh", "-c"] + extra_args_list[1:]
+        container_exec_env = _make_runtime_env(global_context)
+        if global_context.verbose:
+            print(f"Running compose exec {service_name} {command_to_exec}")
+        try:
+            ctx.obj.docker.compose.execute(service_name, command_to_exec, envs=container_exec_env)
+        except DockerException as error:
+            print(f"container command returned error exit status")
+
+
+@command.command()
+@click.argument('extra_args', nargs=-1)  # help: command: logs <service1> <service2>
+@click.pass_context
+def logs(ctx, extra_args):
+    global_context = ctx.parent.parent.obj
+    extra_args_list = list(extra_args) or None
+    if not global_context.dry_run:
+        if global_context.verbose:
+            print("Running compose logs")
+        logs_output = ctx.obj.docker.compose.logs(services=extra_args_list if extra_args_list is not None else [])
+        print(logs_output)
 
 
 def get_stack_status(ctx, stack):
@@ -137,7 +175,7 @@ def get_stack_status(ctx, stack):
     ctx_copy = copy.copy(ctx)
     ctx_copy.stack = stack
 
-    cluster_context = _make_cluster_context(ctx_copy, None, None, None)
+    cluster_context = _make_cluster_context(ctx_copy, None, None, None, None)
     docker = DockerClient(compose_files=cluster_context.compose_files, compose_project_name=cluster_context.cluster)
     # TODO: refactor to avoid duplicating this code above
     if ctx.verbose:
@@ -162,7 +200,7 @@ def _make_runtime_env(ctx):
     return container_exec_env
 
 
-def _make_cluster_context(ctx, include, exclude, cluster):
+def _make_cluster_context(ctx, include, exclude, cluster, env_file):
 
     if ctx.local_stack:
         dev_root_path = os.getcwd()[0:os.getcwd().rindex("stack-orchestrator")]
@@ -235,16 +273,17 @@ def _make_cluster_context(ctx, include, exclude, cluster):
     if ctx.verbose:
         print(f"files: {compose_files}")
 
-    return cluster_context(cluster, compose_files, pre_start_commands, post_start_commands, cluster_config)
+    return cluster_context(cluster, compose_files, pre_start_commands, post_start_commands, cluster_config, env_file)
 
 
 class cluster_context:
-    def __init__(self, cluster, compose_files, pre_start_commands, post_start_commands, config) -> None:
+    def __init__(self, cluster, compose_files, pre_start_commands, post_start_commands, config, env_file) -> None:
         self.cluster = cluster
         self.compose_files = compose_files
         self.pre_start_commands = pre_start_commands
         self.post_start_commands = post_start_commands
         self.config = config
+        self.env_file = env_file
 
 
 def _convert_to_new_format(old_pod_array):

tests/deploy/run-deploy-test.sh

@@ -0,0 +1,48 @@
+#!/usr/bin/env bash
+set -e
+if [ -n "$CERC_SCRIPT_DEBUG" ]; then
+  set -x
+fi
+# Dump environment variables for debugging
+echo "Environment variables:"
+env
+# Test basic stack-orchestrator deploy
+echo "Running stack-orchestrator deploy test"
+# Bit of a hack, test the most recent package
+TEST_TARGET_SO=$( ls -t1 ./package/laconic-so* | head -1 )
+# Set a non-default repo dir
+export CERC_REPO_BASE_DIR=~/stack-orchestrator-test/repo-base-dir
+echo "Testing this package: $TEST_TARGET_SO"
+echo "Test version command"
+reported_version_string=$( $TEST_TARGET_SO version )
+echo "Version reported is: ${reported_version_string}"
+echo "Cloning repositories into: $CERC_REPO_BASE_DIR"
+rm -rf $CERC_REPO_BASE_DIR
+mkdir -p $CERC_REPO_BASE_DIR
+# Test bringing the test container up and down
+# with and without volume removal
+$TEST_TARGET_SO --stack test setup-repositories
+$TEST_TARGET_SO --stack test build-containers
+$TEST_TARGET_SO --stack test deploy up
+$TEST_TARGET_SO --stack test deploy down
+# The next time we bring the container up the volume will be old (from the previous run above)
+$TEST_TARGET_SO --stack test deploy up
+log_output_1=$( $TEST_TARGET_SO --stack test deploy logs )
+if [[ "$log_output_1" == *"Filesystem is old"* ]]; then
+    echo "Retain volumes test: passed"
+else
+    echo "Retain volumes test: FAILED"
+    exit 1
+fi
+$TEST_TARGET_SO --stack test deploy down --delete-volumes
+# Now when we bring the container up the volume will be new again
+$TEST_TARGET_SO --stack test deploy up
+log_output_2=$( $TEST_TARGET_SO --stack test deploy logs )
+if [[ "$log_output_2" == *"Filesystem is fresh"* ]]; then
+    echo "Delete volumes test: passed"
+else
+    echo "Delete volumes test: FAILED"
+    exit 1
+fi
+$TEST_TARGET_SO --stack test deploy down --delete-volumes
+echo "Test passed"