HackMD
How to migrate from lighthouse-docker to eth-docker
Step 1: Stop the lighthouse-docker validator
In order to avoid getting slashed, you need to shutdown and permanently disable the Lighthouse node that is running via lighthouse-docker.
To do that, cd to the lighthouse-docker directory:
(your path may be slightly different)
Then run this command to bring down the running node:
You should see output like this:
Then, to prevent accidentally restarting the node, move the node's data where lighthouse-docker won't find it:
You can verify that lighthouse-docker has been successfully disabled using this command:
You should see:
You should also verify that nothing is running under docker:
Output should be:
With no running processes listed.
Step 2: sign up for an Infura account
In order to make the transition as easy as possible we'll use Infura to checkpoint sync.
Make an Infura account at https://infura.io/ and sign up for an "ETH2" node.
Then make a note of the URL that looks like this:
You'll need that URL in the next step.
Step 3: set up a new node with eth-docker
Clone the eth-docker repository into ~/eth-docker:
Next, run the ethd setup wizard:
Select the following options in order:
- Select Network: Ethereum Mainnet
- Select deployment type: Ethereum node
- Select consensus client: Lighthouse (Rust)
- Select execution client: choose whichever you would like, Geth is fine
- Rapid sync with remote checkpoint, e.g. Infura? Yes
- Configure remote consensus client: put the URL from step 2
- Grafana dashboards? No (unless you want that)
- Configure Rewards Address: any Ethereum/eth1 address of your choice
- Configure Graffiti: whatever you want
The wizard will now finish and output a few logs that end with something like this:
Stop and ask for help if you get anything that looks like an error.
Step 4: import keys to eth-docker
Run this command, where ~/lighthouse-docker is the path to your old lighthouse-docker directory:
Enter the password when prompted. You should see output like the following:
Step 5: check validator is offline, then start
Check your validator indices on https://beaconcha.in to confirm that they're currently offline. They should have been offline and failing to attest since you completed Step 1. Ensure that at least 15 minutes have elapsed since you started.
IF YOUR VALIDATORS ARE NOT OFFLINE AT THIS POINT YOU WILL GET SLASHED. IF YOU'RE UNSURE STOP RIGHT NOW AND SEEK HELP.
Heeding the above warning, start the new node:
You can check on the logs with:
Which should show:
If you instead see errors from Infura just try re-running the log command to see if succeeds on a retry. The error case looks like this:
Check on the validator with ./ethd logs --tail 200 -f validator. It should show:
After a few minutes you should see that your beacon node is synced and your validator is attesting again. Check https://beaconcha.in again to confirm attestations are being processed by the rest of the network.
Troubleshooting
If you run into issues please ping @sproul on Discord or ask for help on the Lighthouse server, or the lighthouse-helpdesk channel on the EthStaker server. The best time to ping me is 9am-5pm Australia eastern standard time (approx UTC+10).
You may also find the docs for eth-docker helpful:
Cleaning Up
Once you're happy with your new eth-docker setup, and confident that you have a backup of your keys somewhere, you can delete the old lighthouse-docker data to save space:
