bug: lightpush fails with `Failed to request a message push: dial_failure` after the peer node restart

To reproduce

Start relay node and subscribe to a topic
Start lightpush node and connect it to the node above
Check that lightpush works
Restart relay node and re-subscribe to a topic
Check that lightpush works

Expected behavior

Should work

Actual behavior

Failed to request a message push: dial_failure

Script to reproduce it:

#!/bin/bash
printf "\nAssuming you already have a docker network called waku\n"
# if not something like this should create it: docker network create --driver bridge --subnet 172.18.0.0/16 --gateway 172.18.0.1 waku

cluster_id=2
pubsub_topic="/waku/2/rs/$cluster_id/0"
node_1=harbor.status.im/wakuorg/nwaku:latest
node_2=harbor.status.im/wakuorg/nwaku:latest
ext_ip="172.18.204.9"
tcp_port="37344"

printf "\nStarting containers\n"

container_id1=$(docker run -d -i -t -p 37343:37343 -p $tcp_port:$tcp_port -p 37345:37345 -p 37346:37346 -p 37347:37347 $node_1 --listen-address=0.0.0.0 --rest=true --rest-admin=true --websocket-support=true --log-level=TRACE --rest-relay-cache-capacity=100 --websocket-port=37345 --rest-port=37343 --tcp-port=$tcp_port --discv5-udp-port=37346 --rest-address=0.0.0.0 --nat=extip:$ext_ip --peer-exchange=true --discv5-discovery=true --cluster-id=$cluster_id --metrics-server=true --metrics-server-address=0.0.0.0 --metrics-server-port=37347 --metrics-logging=true --pubsub-topic=/waku/2/rs/2/0 --lightpush=true --relay=true)
docker network connect --ip $ext_ip waku $container_id1

printf "\nSleeping 2 seconds\n"
sleep 2

response=$(curl -X GET "http://127.0.0.1:37343/debug/v1/info" -H "accept: application/json")
enrUri=$(echo $response | jq -r '.enrUri')

# Extract the first non-WebSocket address
ws_address=$(echo $response | jq -r '.listenAddresses[] | select(contains("/ws") | not)')

# Check if we got an address, and construct the new address with it
if [[ $ws_address != "" ]]; then
    identifier=$(echo $ws_address | awk -F'/p2p/' '{print $2}')
    if [[ $identifier != "" ]]; then
        multiaddr_with_id="/ip4/${ext_ip}/tcp/${tcp_port}/p2p/${identifier}"
        echo $multiaddr_with_id
    else
        echo "No identifier found in the address."
        exit 1
    fi
else
    echo "No non-WebSocket address found."
    exit 1
fi

container_id2=$(docker run -d -i -t -p 25908:25908 -p 25909:25909 -p 25910:25910 -p 25911:25911 -p 25912:25912 $node_2 --listen-address=0.0.0.0 --rest=true --rest-admin=true --websocket-support=true --log-level=TRACE --rest-relay-cache-capacity=100 --websocket-port=25910 --rest-port=25908 --tcp-port=25909 --discv5-udp-port=25911 --rest-address=0.0.0.0 --nat=extip:172.18.141.214 --peer-exchange=true --discv5-discovery=true --cluster-id=$cluster_id --pubsub-topic=/waku/2/rs/2/0 --lightpush=true --relay=false --discv5-bootstrap-node=$enrUri --lightpushnode=$multiaddr_with_id)

docker network connect --ip 172.18.141.214 waku $container_id2

printf "\nSleeping 10 seconds\n"
sleep 10

printf "\nSubscribe\n"
curl -v -X POST "http://127.0.0.1:37343/relay/v1/subscriptions" -H "Content-Type: application/json" -d '["/waku/2/rs/2/0"]'

printf "\nSleeping 2 seconds\n"
sleep 2

printf "\nLightpush message on subscribed pubusub topic\n"                            
curl -v -X POST "http://127.0.0.1:25908/lightpush/v1/message" -H "Content-Type: application/json" -d '{"pubsubTopic": "/waku/2/rs/2/0", "message": {"payload": "TGlnaHQgcHVzaCB3b3JrcyEh", "contentTopic": "/myapp/1/latest/proto", "timestamp": 1712149720320589312}}'

printf "\nRestarting NODE 1\n"  
docker restart $container_id1

printf "\nSleeping 10 seconds\n"
sleep 10

printf "\nSubscribe\n"
curl -v -X POST "http://127.0.0.1:37343/relay/v1/subscriptions" -H "Content-Type: application/json" -d '["/waku/2/rs/2/0"]'

printf "\nSleeping 2 seconds\n"
sleep 2

printf "\nLightpush message on subscribed pubusub topic\n"                            
curl -v -X POST "http://127.0.0.1:25908/lightpush/v1/message" -H "Content-Type: application/json" -d '{"pubsubTopic": "/waku/2/rs/2/0", "message": {"payload": "TGlnaHQgcHVzaCB3b3JrcyEh", "contentTopic": "/myapp/1/latest/proto", "timestamp": 1712149720320589312}}'

Logs lightpush_node.log relay_node.log

This error happens because when we restart the lightpush service node container (container_id1) a new multiaddress is generated, so the lightpush client node is trying to dial the old multiaddress with no response.

To fix that, we have to start the service node with the --nodekey parameter. For example, using '--nodekey=6a29e767c96a2a380bb66b9a6ffcd6eb54049e14d796a1d866307b8beb7aee58'

That is, replacing line 15 in the script for

container_id1=$(docker run -d -i -t -p 37343:37343 -p $tcp_port:$tcp_port -p 37345:37345 -p 37346:37346 -p 37347:37347 $node_1 --listen-address=0.0.0.0 --rest=true --rest-admin=true --websocket-support=true --log-level=TRACE --rest-relay-cache-capacity=100 --websocket-port=37345 --rest-port=37343 --tcp-port=$tcp_port --discv5-udp-port=37346 --rest-address=0.0.0.0 --nat=extip:$ext_ip --peer-exchange=true --discv5-discovery=true --cluster-id=$cluster_id --metrics-server=true --metrics-server-address=0.0.0.0 --metrics-server-port=37347 --metrics-logging=true --pubsub-topic=/waku/2/rs/2/0 --lightpush=true --relay=true --nodekey=6a29e767c96a2a380bb66b9a6ffcd6eb54049e14d796a1d866307b8beb7aee58)

We avoid having a new multiaddress generated after container restart, and there's no more dial failures anymore

@fbarbu15 please confirm if it makes sense and works for you too

waku-org / nwaku