Restart support for fixturenet-lotus (#499)

* Use ip utility to get the required miner node multiaddr

* Persist lotus node data to support restarts

* Add clean up steps to instructions

* Fix lotus-seed sector-dir arg

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

@@ -11,6 +11,7 @@ services:
       - ../config/fixturenet-lotus/fund-account.sh:/fund-account.sh
       - lotus_miner_params:/var/tmp/filecoin-proof-parameters
       - lotus-shared:/root/.lotus-shared
+      - lotus_miner_data:/root/data
     entrypoint: ["sh", "/docker-entrypoint-scripts.d/setup-miner.sh"]
     ports:
       - "1234"
@@ -30,6 +31,7 @@ services:
       - ../config/fixturenet-lotus/setup-node.sh:/docker-entrypoint-scripts.d/setup-node.sh
       - lotus_node_1_params:/var/tmp/filecoin-proof-parameters
       - lotus-shared:/root/.lotus-shared
+      - lotus_node_1_data:/root/data
     healthcheck:
       test: ["CMD", "nc", "-vz", "localhost", "1234"]
       interval: 30s
@@ -57,6 +59,7 @@ services:
       - ../config/fixturenet-lotus/setup-node.sh:/docker-entrypoint-scripts.d/setup-node.sh
       - lotus_node_2_params:/var/tmp/filecoin-proof-parameters
       - lotus-shared:/root/.lotus-shared
+      - lotus_node_2_data:/root/data
     healthcheck:
       test: ["CMD", "nc", "-vz", "localhost", "1234"]
       interval: 30s
@@ -77,3 +80,6 @@ volumes:
   lotus_node_1_params:
   lotus_node_2_params:
   lotus-shared:
+  lotus_miner_data:
+  lotus_node_1_data:
+  lotus_node_2_data:

app/data/config/fixturenet-lotus/lotus-env.env

@@ -1,5 +1,5 @@
-LOTUS_PATH=/root/.lotus-local-net
-LOTUS_MINER_PATH=/root/.lotus-miner-local-net
+LOTUS_PATH=/root/data/.lotus-local-net
+LOTUS_MINER_PATH=/root/data/.lotus-miner-local-net
 LOTUS_SKIP_GENESIS_CHECK=_yes_
 LOTUS_FEVM_ENABLEETHRPC=true
 CGO_CFLAGS_ALLOW="-D__BLST_PORTABLE__"

app/data/config/fixturenet-lotus/setup-miner.sh

@@ -15,12 +15,16 @@ else
   echo "Existing proof params found"
 fi
 
-lotus-seed pre-seal --sector-size 2KiB --num-sectors 2
-lotus-seed genesis new localnet.json
-lotus-seed genesis add-miner localnet.json ~/.genesis-sectors/pre-seal-t01000.json
+# if genesis is not already setup
+if [ ! -f /root/data/localnet.json ]; then
+  lotus-seed --sector-dir /root/data/.genesis-sectors pre-seal --sector-size 2KiB --num-sectors 2
+  lotus-seed --sector-dir /root/data/.genesis-sectors genesis new /root/data/localnet.json
+  lotus-seed --sector-dir /root/data/.genesis-sectors genesis add-miner /root/data/localnet.json /root/data/.genesis-sectors/pre-seal-t01000.json
+fi
 
 # start daemon
-nohup lotus daemon --lotus-make-genesis=devgen.car --profile=bootstrapper --genesis-template=localnet.json --bootstrap=false > /var/log/lotus.log 2>&1 &
+# /root/.lotus-shared/devgen.car path
+nohup lotus daemon --lotus-make-genesis=/root/.lotus-shared/devgen.car --profile=bootstrapper --genesis-template=/root/data/localnet.json --bootstrap=false > /var/log/lotus.log 2>&1 &
 
 # Loop until the daemon is started
 echo "Waiting for daemon to start..."
@@ -29,22 +33,18 @@ while ! grep -q "started ChainNotify channel" /var/log/lotus.log ; do
 done
 echo "Daemon started."
 
-# copy genesis file to shared volume
-cp /devgen.car /root/.lotus-shared
-
 # publish bootnode peer info to shared volume
-# TODO: Improve exporting public address to shared volume
-lotus net listen | awk 'NR==4{print}' > /root/.lotus-shared/miner.addr
+lotus net listen | grep "$(ip addr | grep inet | grep -v '127.0.0.1' | sort | head -1 | awk '{print $2}' | cut -d '/' -f1)" | head -1 > /root/.lotus-shared/miner.addr
 
 # if miner not already initialized
 if [ ! -d $LOTUS_MINER_PATH ]; then
   # initialize miner
-  lotus wallet import --as-default ~/.genesis-sectors/pre-seal-t01000.key
+  lotus wallet import --as-default /root/data/.genesis-sectors/pre-seal-t01000.key
 
   # fund a known account for usage
   /fund-account.sh
 
-  lotus-miner init --genesis-miner --actor=t01000 --sector-size=2KiB --pre-sealed-sectors=~/.genesis-sectors --pre-sealed-metadata=~/.genesis-sectors/pre-seal-t01000.json --nosync
+  lotus-miner init --genesis-miner --actor=t01000 --sector-size=2KiB --pre-sealed-sectors=/root/data/.genesis-sectors --pre-sealed-metadata=/root/data/.genesis-sectors/pre-seal-t01000.json --nosync
 fi
 
 # start miner

app/data/config/fixturenet-lotus/setup-node.sh

@@ -9,20 +9,17 @@ while [ ! -f /root/.lotus-shared/miner.addr ]; do
 done
 echo "Resuming..."
 
-# if not already initialized
-if [ ! -f $LOTUS_PATH/config.toml ]; then
-  # init node config
-  mkdir $LOTUS_PATH
-  lotus config default > $LOTUS_PATH/config.toml
+# init node config
+mkdir -p $LOTUS_PATH
+lotus config default > $LOTUS_PATH/config.toml
 
-  # add bootstrap peer info if available
-  if [ -f /root/.lotus-shared/miner.addr ]; then
-    MINER_ADDR=\"$(cat /root/.lotus-shared/miner.addr)\"
-    # add bootstrap peer id to config file
-    sed -i "/^\[Libp2p\]/a \ \ BootstrapPeers = [$MINER_ADDR]" $LOTUS_PATH/config.toml
-  else
-    echo "Bootstrap peer info not found, unable to configure. Manual peering will be required."
-  fi
+# add bootstrap peer info if available
+if [ -f /root/.lotus-shared/miner.addr ]; then
+  MINER_ADDR=\"$(cat /root/.lotus-shared/miner.addr)\"
+  # add bootstrap peer id to config file
+  sed -i "/^\[Libp2p\]/a \ \ BootstrapPeers = [$MINER_ADDR]" $LOTUS_PATH/config.toml
+else
+  echo "Bootstrap peer info not found, unable to configure. Manual peering will be required."
 fi
 
 # start node

app/data/container-build/cerc-lotus/Dockerfile

@@ -99,7 +99,7 @@ CMD ["-help"]
 FROM lotus-base AS lotus-all-in-one
 
 # Install netcat for healthcheck
-RUN apt-get update && apt-get install -y netcat
+RUN apt-get update && apt-get install -y netcat && apt-get install -y iproute2
 
 ENV FILECOIN_PARAMETER_CACHE /var/tmp/filecoin-proof-parameters
 ENV LOTUS_MINER_PATH /var/lib/lotus-miner

app/data/stacks/azimuth/README.md

@@ -55,7 +55,7 @@ This should create the required docker images in the local image registry.
 
 ## Clean up
 
-Stop all the services running in background run:
+Stop all the services running in background:
 
 ```bash
 laconic-so --stack azimuth deploy-system down

app/data/stacks/fixturenet-lotus/README.md

@@ -14,6 +14,9 @@ $ laconic-so --stack fixturenet-lotus build-containers
 ```
 $ laconic-so --stack fixturenet-lotus deploy --cluster lotus up
 ```
+
+Note: When running for the first time (or after clean up), the services will take some time to start properly as the Lotus nodes download the proof params (which are persisted to volumes)
+
 Correct operation should be verified by checking the container logs with:
 ```
 $ laconic-so --stack fixturenet-lotus deploy --cluster lotus logs lotus-miner
@@ -26,3 +29,19 @@ $ laconic-so --stack fixturenet-lotus deploy --cluster lotus exec lotus-miner "l
 $ laconic-so --stack fixturenet-lotus deploy --cluster lotus exec lotus-node-1 "lotus status"
 $ laconic-so --stack fixturenet-lotus deploy --cluster lotus exec lotus-node-2 "lotus status"
 ```
+
+## 4. Clean up
+
+Stop all the services running in background:
+```
+$ laconic-so --stack fixturenet-lotus deploy --cluster lotus down
+```
+
+Clear volumes created by this stack:
+```
+# List all relevant volumes
+$ docker volume ls -q --filter "name=lotus"
+
+# Remove all the listed volumes
+$ docker volume rm $(docker volume ls -q --filter "name=lotus")
+```

app/data/stacks/mainnet-eth/README.md

@@ -70,12 +70,41 @@ To permanently *delete* the stack's data volumes run:
 $ laconic-so deployment --dir mainnet-eth-deployment stop --delete-data-volumes
 ```
 After deleting the volumes, any subsequent re-start will begin chain sync from cold.
+
+## Ports
+It is usually necessary to expose certain container ports on one or more the host's addresses to allow incoming connections.
+Any ports defined in the Docker compose file are exposed by default with random port assignments, but the values can be
+customized by editing the "spec" file generated by `laconic-so deploy init`.
+
+In this example, ports `8545` and `5052` have been assigned to a specific addresses/port combination on the host, while
+port `40000` has been left with random assignment:
+```
+$ cat mainnet-eth-spec.yml
+stack: mainnet-eth
+ports:
+  mainnet-eth-geth-1:
+   - '10.10.10.10:8545:8545'
+   - '40000'
+  mainnet-eth-lighthouse-1:
+   - '10.10.10.10:5052:5052'
+volumes:
+  mainnet_eth_config_data: ./data/mainnet_eth_config_data
+  mainnet_eth_geth_1_data: ./data/mainnet_eth_geth_1_data
+  mainnet_eth_lighthouse_1_data: ./data/mainnet_eth_lighthouse_1_data
+```
+
 ## Data volumes
 Container data volumes are bind-mounted to specified paths in the host filesystem.
 The default setup (generated by `laconic-so deploy init`) places the volumes in the `./data` subdirectory of the deployment directory:
 ```
 $ cat mainnet-eth-spec.yml
 stack: mainnet-eth
+ports:
+  mainnet-eth-geth-1:
+   - '10.10.10.10:8545:8545'
+   - '40000'
+  mainnet-eth-lighthouse-1:
+   - '10.10.10.10:5052:5052'
 volumes:
   mainnet_eth_config_data: ./data/mainnet_eth_config_data
   mainnet_eth_geth_1_data: ./data/mainnet_eth_geth_1_data

app/data/stacks/mobymask-v2/README.md

@@ -93,7 +93,7 @@ Follow the [demo](./demo.md) to try out the MobyMask app with L2 chain
 
 ## Clean up
 
-Stop all the services running in background run:
+Stop all the services running in background:
 
 ```bash
 laconic-so --stack mobymask-v2 deploy --cluster mobymask_v2 down 30

app/data/stacks/sushiswap/README.md

@@ -22,13 +22,15 @@ Deploy the stack:
 laconic-so --stack sushiswap deploy --cluster sushiswap up
 ```
 
+Note: When running for the first time (or after clean up), the services will take some time to start as Lotus nodes in the fixturenet download the proof params
+
 ## Tests
 
 Follow [smoke-tests.md](./smoke-tests.md) to run smoke tests
 
 ## Clean up
 
-Stop all the services running in background run:
+Stop all the services running in background:
 
 ```bash
 laconic-so --stack sushiswap deploy --cluster sushiswap down

app/deployment_create.py

@@ -27,6 +27,20 @@ from app.deploy_types import DeploymentContext, DeployCommandContext
 def _make_default_deployment_dir():
     return "deployment-001"
 
+def _get_ports(stack):
+    ports = {}
+    parsed_stack = get_parsed_stack_config(stack)
+    pods = parsed_stack["pods"]
+    yaml = get_yaml()
+    for pod in pods:
+        pod_file_path = os.path.join(get_compose_file_dir(), f"docker-compose-{pod}.yml")
+        parsed_pod_file = yaml.load(open(pod_file_path, "r"))
+        if "services" in parsed_pod_file:
+            for svc_name, svc in parsed_pod_file["services"].items():
+                if "ports" in svc:
+                    # Ports can appear as strings or numbers.  We normalize them as strings.
+                    ports[svc_name] = [ str(x) for x in svc["ports"] ]
+    return ports
 
 def _get_named_volumes(stack):
     # Parse the compose files looking for named volumes
@@ -62,24 +76,30 @@ def _create_bind_dir_if_relative(volume, path_string, compose_dir):
 
 # See: https://stackoverflow.com/questions/45699189/editing-docker-compose-yml-with-pyyaml
 def _fixup_pod_file(pod, spec, compose_dir):
-    # Fix up volumes
-    if "volumes" in spec:
-        spec_volumes = spec["volumes"]
-        if "volumes" in pod:
-            pod_volumes = pod["volumes"]
-            for volume in pod_volumes.keys():
-                if volume in spec_volumes:
-                    volume_spec = spec_volumes[volume]
-                    volume_spec_fixedup = volume_spec if Path(volume_spec).is_absolute() else f".{volume_spec}"
-                    _create_bind_dir_if_relative(volume, volume_spec, compose_dir)
-                    new_volume_spec = {"driver": "local",
-                                       "driver_opts": {
-                                           "type": "none",
-                                           "device": volume_spec_fixedup,
-                                           "o": "bind"
-                                        }
+   # Fix up volumes
+   if "volumes" in spec:
+       spec_volumes = spec["volumes"]
+       if "volumes" in pod:
+           pod_volumes = pod["volumes"]
+           for volume in pod_volumes.keys():
+               if volume in spec_volumes:
+                   volume_spec = spec_volumes[volume]
+                   volume_spec_fixedup = volume_spec if Path(volume_spec).is_absolute() else f".{volume_spec}"
+                   _create_bind_dir_if_relative(volume, volume_spec, compose_dir)
+                   new_volume_spec = {"driver": "local",
+                                      "driver_opts": {
+                                          "type": "none",
+                                          "device": volume_spec_fixedup,
+                                          "o": "bind"
                                        }
-                    pod["volumes"][volume] = new_volume_spec
+                                      }
+                   pod["volumes"][volume] = new_volume_spec
+   # Fix up ports
+   if "ports" in spec:
+       spec_ports = spec["ports"]
+       for container_name, container_ports in spec_ports.items():
+           if container_name in pod["services"]:
+               pod["services"][container_name]["ports"] = container_ports
 
 
 def call_stack_deploy_init(deploy_command_context):
@@ -148,12 +168,18 @@ def init(ctx, output):
         spec_file_content.update(default_spec_file_content)
     if verbose:
         print(f"Creating spec file for stack: {stack}")
+
+    ports = _get_ports(stack)
+    if ports:
+        spec_file_content["ports"] = ports
+
     named_volumes = _get_named_volumes(stack)
     if named_volumes:
         volume_descriptors = {}
         for named_volume in named_volumes:
             volume_descriptors[named_volume] = f"./data/{named_volume}"
         spec_file_content["volumes"] = volume_descriptors
+
     with open(output, "w") as output_file:
         yaml.dump(spec_file_content, output_file)