helium validator + watchtower (with consensus check) + exporter

docker-compose.yml

# launch: docker-compose up -d
version: "3.7"
services:
  validator:
    image: quay.io/team-helium/validator:latest-validator-amd64
    container_name: validator
    init: true
    restart: always
    ports:
    - "2154:2154"
    - "127.0.0.1:4467:4467" # jsonrpc
    volumes:
    - "${HOME}/validator_data/:/var/data"
    labels:
      "com.centurylinklabs.watchtower.lifecycle.pre-update": >
        sh -c 'while [ "$$(miner info in_consensus)" == "true" ]; do echo "in consensus, loop-waiting"; sleep 30; done'
      "com.centurylinklabs.watchtower.lifecycle.pre-update-timeout": 43200 # 12hrs should be longer than a validator is in the consensus group.
  val_watchtower:
    image: containrrr/watchtower:latest
    container_name: val_watchtower
    restart: always
    volumes:
    - /var/run/docker.sock:/var/run/docker.sock
    command: validator miner_exporter # the container name(s) to monitor
    environment:
    - WATCHTOWER_POLL_INTERVAL=21600 # check for new image every 6 hrs
    - WATCHTOWER_LIFECYCLE_HOOKS=true
    #- WATCHTOWER_DEBUG=true # uncomment to see watchtower waiting for consensus
  miner_exporter:
    image: ghcr.io/tedder/miner_exporter:latest
    container_name: miner_exporter
    restart: always
    volumes:
    - /var/run/docker.sock:/var/run/docker.sock
    ports:
    - "127.0.0.1:9825:9825"

🙏

Here's a docker run version for watchtower with the in_consensus pre-update check. (waiting to test in the wild...)

NOTE: This needs tweaking. Watchtower restarts and restarts....

docker run -d \
--restart always \
--name watchtower \
-v /var/run/docker.sock:/var/run/docker.sock \
containrrr/watchtower validator miner_exporter \
--cleanup \
--label=com.centurylinklabs.watchtower.lifecycle.pre-update="sh -c 'while [ "\$\(miner info in_consensus\)" == "true" ]; do echo "in consensus, loop-waiting"; sleep 30; done'" \
--label=com.centurylinklabs.watchtower.lifecycle.pre-update-timeout=43200 \
-e WATCHTOWER_LIFECYCLE_HOOKS=true \
--interval 21600

using the json rpc does work in 0.73, not sure it's the best solution or how to integrate it, but thought I'd share. Also I don't like the dep. on jq

curl -X POST -H 'Content-Type: application/json' -d '{"jsonrpc":"2.0","id":1,"method":"info_in_consensus", "params":[]}' http://127.0.0.1:4467 | jq ".result.in_consensus"

I think you are going to need a "restart: always" in each service section so that everything will start on reboot?

Thanks for writing this docker-compose.yml file. I believe the volume mount for 'validator' container should be

    volumes:
    - "${HOME}/validator_data:/var/data"

@philltran removing a trailing slash? It should work either way.

You are right. The issue I was having was with the quay tag. I needed to use validator for the main net and not val for testnet.

good call! I was having trouble mentally diffing it.

@tedder Don't you need a "restart: always" in each service section so that everything will start on reboot?

@jkowalcz not a bad idea. added.

Might want to add 8080 port for light hotspots

https://gist.github.com/rootulp/4fe3a7e420237a5632f3201faa246ab4 has two editions to this docker-compose.yml

1. Expose 8080 per deployment guide
2. Increase ulimits to avoid No file descriptors available error