For the Teledisko DAO graph-node configuration, check ./provision.

Connect to the production server

To manage containers, update the subgraph: ssh [email protected]

To admin the server: ssh [email protected]

Troubleshooting

Deploy a new subgraph

Make sure you've updated all contract addresses (if #9 has not been implemented) in the dao/subgraph.yaml file. Update startBlock if needed.

Connect to [email protected], run tmux attach to connect to the current tmux session, go to /home/worker/subgraph/dao, run

• staging instance: pnpm run remove-local ; pnpm run create-local ; pnpm deploy-staging
• production instance: pnpm run remove-local ; pnpm run create-local ; pnpm deploy-production

Contract changed? Update and regenerate ABI

TBD

ERRO the genesis block hash for chain tevmos has changed from X to Y since the last time we ran, component: BlockStore

See: graphprotocol/graph-node#3655

Solution: stop docker compose, remove /home/worker/subgraph/data dir, restart docker compose.

ERRO Connection to provider failed when connecting to TEVMOS

See: graphprotocol/graph-node#3699

Solution:

• check if the testnet has been restarted and recreated using a new genesis file: https://github.com/evmos/testnets
• if so, check the new lowest block number
• update the lowest block number in docker-compose.yml, env variable GRAPH_ETHEREUM_GENESIS_BLOCK_NUMBER

It doesn't index

I (Alberto) think it's related to this log message INFO Scanning blocks [0, 0], range_size: 1 [...]. I noticed that in testnet the genesis block changes (do they cleanup the chain every now and then?). Anyway in our .env file we have the entry GENESIS_BLOCK=<number>, it seems to me that if this number is wrong (because the chain has been cleaned up, and it doesn't exist anymore) then the indexing doesn't work. What you should do in that case is check whenver the genesis block changed (I did a binary search in their explorer), update the env, restart docker compose.

Installation

npm install -g @graphprotocol/graph-cli

Local development

Setup

# make sure your env contains the following variable
ETHEREUM=<tevmos:https://eth.bd.evmos.dev:8545 | evmos:https://eth.bd.evmos.org:8545>
GENESIS_BLOCK=<genesis block>

# install and run ethnode or ganache-cli
# deploy contracts locally

git clone https://github.com/graphprotocol/graph-node/
cd graph-node/docker
docker-compose up

Deploy locally

graph create --node http://localhost:8020/ <subgraph-name>
graph init --product hosted-service --from-contract <address> --allow-simple-name --index-events --contract-name <contract-name> --abi <dir> --network mainnet --protocol ethereum <subgraph-name>
graph deploy <subgraph-name> -g http://localhost:8020/ -i http://localhost:5001/

Use hosted service (aka staging)

Setup

# deploy the contract to rinkeby, remember the transaction id
# go to https://thegraph.com/hosted-service/ in the dashboard
# select the TelediskoDAO account
# add a new subgraph if necessary
# grab the Access Token and authenticate
 
graph auth --product hosted-service $ACCESS_TOKEN

Deploy

graph init --product hosted-service --from-contract <address> --allow-simple-name --index-events --contract-name <contract-name> --abi <dir> --network rinkeby --protocol ethereum TelediskoDAO/<subgraph-name>
cd <subgraph-name>
# visit rinkeby.etherscan.io and get the block number of the contract transaction id
# open "subgraph.yml" and under the "source" node add "startBlock: <block number>"
graph deploy --product hosted-service TelediskoDAO/<subgraph-name>

You should now see the indexed graph on the dashboard.

Dumb backup

Export:

export IPFS_PATH=subgraph/data/ipfs/

for cid in $(curl 'https://graph.dao.teledisko.com/subgraphs/name/TelediskoDAO/dao' -X POST -H 'Accept: application/json, multipart/mixed' -H 'content-type: application/json' --data-raw '{"query":"query MyQuery {\n  resolutions {\n    ipfsDataURI\n  }\n}","variables":null,"operationName":"MyQuery","extensions":{"headers":null}}' | jq ".data.resolutions[].ipfsDataURI" -r)
do
    ipfs cat ${cid} > backup/${cid}
done

Restore:

ipfs add *