Skip to main content

AVS startup scripts

How to start an AVS in bridge mode​

At the moment Ivynet Client can only track metrics from AVSs started in the default (bridge) docker network mode. Not every AVS has documentation on how to set up these ports, as some start in host network mode. Below are a few examples on how to modify a startup script/procedure to start a container with metrics ports exposed.

K3​

The K3 AVS exposes metrics on port 4000, so the script to start it can look like:

#!/bin/bash

# Change the IMAGE for the mainnet
VERSION=latest
IMAGE=k3official/k3-labs-avs-operator-dev:{VERSION}
NAME=${1:-k3}

echo "Kill docker container:"
docker ps | grep -q ${NAME} && docker kill ${NAME}
echo "Remove docker container:"
docker ps -a | grep -q ${NAME} && docker rm ${NAME}

echo "=========="
echo "Pull the image:"
docker pull ${IMAGE}

echo "=========="
echo "Run the container:"
echo "${NAME}"
docker run \
  -d \
  --env-file .env \
  --name ${NAME} \
  -p 3000:3000 \
  -p 4001:4001 \
  -p 4002:4002 \
  -p 9876:9876 \
  ${IMAGE} performer

The script not only starts the AVS, exposes all used ports, and names the container (default is 'k3'), but also sends it to the background. The logs can be accessed with the docker logs command, e.g. if the default name is used:

docker logs k3

Predicate​

The Predicate AVS exposes metrics on port 9090, which is also used by other AVSs (e.g. EigenDA). Therefore the script not only exposes the ports used by the program but also changes them to avoid conflicts.

#!/bin/bash

VERSION=1.0.2
IMAGE=ghcr.io/predicatelabs/operator:${VERSION}
NAME=${1:-predicate}

export ENV_FILE=./.env
. $ENV_FILE

echo "Kill docker container:"
docker ps | grep -q ${NAME} && docker kill ${NAME}
echo "Remove docker container:"
docker ps -a | grep -q ${NAME} && docker rm ${NAME}

echo "=========="
echo "Pull the image:"
docker pull ${IMAGE}

echo "=========="
echo "Run the container:"
echo "${NAME}"
docker run \
  -d \
  -p 19010:9010 \
  -p 19090:9090 \
  -p 19091:9091 \
  --name ${NAME} \
  --env-file ${ENV_FILE} \
  -v "${DB_PATH}:/app/data/" \
  -v "${PREDICATE_SIGNING_PRIVATE_KEY_STORE_PATH}:/app/signingkey.json" \
  ${IMAGE} start \
    --db-path /app/data/ \
    --predicate-signing-private-key-store-path /app/signingkey.json

The script not only starts the AVS, exposes all the ports, and names it (default is 'predicate'), but also sends it to the background. The logs can be accessed with the docker logs command, e.g. if the default name is used:

docker logs predicate

Scripts using Ivynet API​

The script is an extended version of the Predicate startup one. Before the container is starts it gets the latest version of it from the ivynet API. It requires Ivynet credentials (at the moment email and password) to be stored in the .ivynet.env file. Additionally, the curl has to be installed.

#!/bin/bash

# load credentials from the .ivynet.env file
source .ivynet.env

# get the version from the Ivynet API server
VERSION=$(curl -u ${IVYNET_USER}:${IVYNET_PASS} -X 'GET' \
  'https://api1.test.ivynet.dev/info/avs/version/Predicate' \
  -H 'accept: application/json' | \
    jq -r ' .[] | select( .chain | contains("holesky"))| .latest_version')

# set the image
IMAGE=ghcr.io/predicatelabs/operator:${VERSION}
NAME=${1:-predicate}

# ensure the content of the .env is loaded
export ENV_FILE=./.env
. $ENV_FILE

echo "Kill docker container:"
docker ps | grep -q ${NAME} && docker kill ${NAME}
echo "Remove docker container:"
docker ps -a | grep -q ${NAME} && docker rm ${NAME}

echo "=========="
echo "Pull the image:"
docker pull ${IMAGE}

echo "=========="
echo "Run the container:"
echo "${NAME}"
docker run \
  -d \
  -p 19010:9010 \
  -p 19090:9090 \
  -p 19091:9091 \
  --name ${NAME} \
  --env-file ${ENV_FILE} \
  -v "${DB_PATH}:/app/data/" \
  -v "${PREDICATE_SIGNING_PRIVATE_KEY_STORE_PATH}:/app/signingkey.json" \
  ${IMAGE} start \
    --db-path /app/data/ \
    --predicate-signing-private-key-store-path /app/signingkey.json