Skip to content

Commit

Permalink
Switch operator instructions ordering (#1504)
Browse files Browse the repository at this point in the history 
This changes the node operator instructions so that op-geth is executed
before op-node, ensuring that op-node won't fail when executed because
it can't find op-geth.

Also removes the step to manually create the jwt token because op-geth
will do this automatically if no token is found at the configured path.
  • Loading branch information
@piersy
piersy authored Sep 27, 2024
1 parent e675812 commit 225a9e9
Showing 1 changed file with 61 additions and 65 deletions.
126 changes: 61 additions & 65 deletions docs/cel2/l2-operator-guide.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -63,8 +63,8 @@ To set up your node, you'll need to run three services: `op-node`, `op-geth`, an
If you plan to run multiple L2 nodes, you’ll need separate instances of `op-node` and `op-geth` for each node, but you can share a single instance of `eigenda-proxy`. Additionally, you'll need an endpoint for the L1 RPC, which can either be a public node or your own L1 node. For Alfajores, the L1 used is [Holesky](https://chainlist.org/chain/17000).

- [Running EigenDA Proxy](#running-eigenda-proxy)
- [Running op-node](#running-op-node)
- [Running op-geth](#running-op-geth)
- [Running op-node](#running-op-node)

### Running EigenDA Proxy

Expand Down Expand Up @@ -135,63 +135,6 @@ docker run --platform linux/amd64 -d \

You can check the logs running `docker logs -f eigenda-proxy` to ensure the service is running correctly.

### Running op-node

Op-node is not a resource-demanding service. We recommend running it on any modern CPU (amd64 or arm64) with at least 2GB of memory. It is stateless, so it does not require any persistent storage.

1. To run op-node, you can use the container image: `us-west1-docker.pkg.dev/devopsre/celo-blockchain-public/op-node:celo9`. Alternatively, you can clone the [celo-org/optimism repository](https://github.com/celo-org/optimism) and build op-node from source:

```bash
git clone https://github.com/celo-org/optimism.git
cd optimism/op-node
git checkout celo9
make op-node
```

2. Download the rollup config file and generate a JWT secret (this value will also be required for configuring the op-geth client):

```bash
OP_NODE_DIR=/tmp/alfajores/op-node
mkdir -p ${OP_NODE_DIR}
wget https://storage.googleapis.com/cel2-rollup-files/alfajores/rollup.json --output-document=${OP_NODE_DIR}/rollup.json

openssl rand -hex 32 > ${OP_NODE_DIR}/jwt.txt
```

3. Run the container (or the binary if preferred). You can use the following example as a reference. If you're using snap sync mode, you need to add the flag `--syncmode=execution-layer`.

```bash
OP_NODE_IMAGE=us-west1-docker.pkg.dev/devopsre/celo-blockchain-public/op-node:celo9

docker run --platform linux/amd64 -d \
--name op-node \
--network=host \
--restart=always \
-v ${OP_NODE_DIR}:/data \
${OP_NODE_IMAGE} \
op-node \
--l1.trustrpc=true \
--l1=https://ethereum-holesky-rpc.publicnode.com \
--l1.beacon=https://ethereum-holesky-beacon-api.publicnode.com \
--l2=http://localhost:8551 \
--l2.jwt-secret=/data/jwt.txt \
--rollup.load-protocol-versions=true \
--rollup.config=/data/rollup.json \
--verifier.l1-confs=4 \
--rpc.addr=127.0.0.1 \
--rpc.port=9545 \
--p2p.listen.tcp=9222 \
--p2p.listen.udp=9222 \
--p2p.priv.path=/data/op-node_p2p_priv.txt \
--p2p.static=/ip4/35.197.25.52/tcp/9222/p2p/16Uiu2HAmQEdyLRSAVZDr5SqbJ1RnKmNDhtQJcEKmemrVxe4FxKwR,/ip4/34.105.22.4/tcp/9222/p2p/16Uiu2HAm1SZBDSugT5MMu7vBY8auDgfZFNhoDeXPLc9Me5FsAxwT,/ip4/34.83.209.168/tcp/9222/p2p/16Uiu2HAmGJAiUX6HLSo4nLh8T984qxzokwL23cVsYuNZy2SrK7C6,/ip4/34.83.214.149/tcp/9222/p2p/16Uiu2HAmAko2Kr3eAjM7tnshtEhYrxQYfKUvN2kwiygeFoBAoi8S,/ip4/34.169.5.52/tcp/9222/p2p/16Uiu2HAmKc6YKHzYgsjBDaj36uAufxpgZFgrzDqVBt6zTPwdhhJD \
--altda.enabled=true \
--altda.da-server=http://localhost:4242 \
--altda.da-service=true \
--altda.verify-on-read=false
```

You can check the logs running `docker logs -f op-node`. If you start op-node before op-geth, it will shut down after a few seconds if it cannot connect to its corresponding op-geth instance. This is normal behavior. It will run successfully once op-geth is running and it can connect to it.

### Running op-geth

Now, let's move on to the op-geth execution client. It will be responsible for executing transactions and generating blocks. We recommend running op-geth on a machine with 8 modern cores (amd64 or arm64), at least 8GB of memory and 200GB of storage. Feel free to adjust these values based on your specific requirements.
Expand Down Expand Up @@ -350,9 +293,6 @@ Please check in your migration logs that the resulting migration block must look
`
:::

##### Start your L2 Node

Now that we have the migrated chaindata, we can start our L2 node using your own migrated chaindata (the `new-db` path) and the genesis file generated by the migration script. You can also use the `rollup.json` file generated by the migration script for your op-node.
Please continue with [executing op-geth](#executing-op-geth) instructions to start your L2 node.

#### Executing op-geth
Expand All @@ -371,8 +311,6 @@ make geth
```bash
OP_GETH_IMAGE=us-west1-docker.pkg.dev/devopsre/celo-blockchain-public/op-geth:celo8

cp ${OP_NODE_DIR}/jwt.txt ${OP_GETH_DATADIR}/jwt.txt

docker run -d \
--name op-geth \
--network=host \
Expand All @@ -386,7 +324,7 @@ docker run -d \
--port=30303 \
--authrpc.addr=127.0.0.1 \
--authrpc.port=8551 \
--authrpc.jwtsecret=/datadir/jwt.txt \
--authrpc.jwtsecret=/datadir/jwt.hex \
--authrpc.vhosts='*' \
--http \
--http.addr=127.0.0.1 \
Expand All @@ -400,7 +338,65 @@ docker run -d \

To see the logs, you can run `docker logs -f op-geth`.

#### Configuring a HistoricalRPCService
### Running op-node

Op-node is not a resource-demanding service. We recommend running it on any modern CPU (amd64 or arm64) with at least 2GB of memory. It is stateless, so it does not require any persistent storage.

1. To run op-node, you can use the container image: `us-west1-docker.pkg.dev/devopsre/celo-blockchain-public/op-node:celo9`. Alternatively, you can clone the [celo-org/optimism repository](https://github.com/celo-org/optimism) and build op-node from source:

```bash
git clone https://github.com/celo-org/optimism.git
cd optimism/op-node
git checkout celo9
make op-node
```

2. Make a directory for op node config and download the rollup config file:

```bash
OP_NODE_DIR=/tmp/alfajores/op-node
mkdir -p ${OP_NODE_DIR}
wget https://storage.googleapis.com/cel2-rollup-files/alfajores/rollup.json --output-document=${OP_NODE_DIR}/rollup.json
```

3. Run the container (or the binary if preferred). You can use the following example as a reference. If you're using snap sync mode, you need to add the flag `--syncmode=execution-layer`.

```bash
OP_NODE_IMAGE=us-west1-docker.pkg.dev/devopsre/celo-blockchain-public/op-node:celo9

cp ${OP_GETH_DATADIR}/jwt.hex ${OP_NODE_DIR}/jwt.hex

docker run --platform linux/amd64 -d \
--name op-node \
--network=host \
--restart=always \
-v ${OP_NODE_DIR}:/data \
${OP_NODE_IMAGE} \
op-node \
--l1.trustrpc=true \
--l1=https://ethereum-holesky-rpc.publicnode.com \
--l1.beacon=https://ethereum-holesky-beacon-api.publicnode.com \
--l2=http://localhost:8551 \
--l2.jwt-secret=/data/jwt.hex \
--rollup.load-protocol-versions=true \
--rollup.config=/data/rollup.json \
--verifier.l1-confs=4 \
--rpc.addr=127.0.0.1 \
--rpc.port=9545 \
--p2p.listen.tcp=9222 \
--p2p.listen.udp=9222 \
--p2p.priv.path=/data/op-node_p2p_priv.txt \
--p2p.static=/ip4/35.197.25.52/tcp/9222/p2p/16Uiu2HAmQEdyLRSAVZDr5SqbJ1RnKmNDhtQJcEKmemrVxe4FxKwR,/ip4/34.105.22.4/tcp/9222/p2p/16Uiu2HAm1SZBDSugT5MMu7vBY8auDgfZFNhoDeXPLc9Me5FsAxwT,/ip4/34.83.209.168/tcp/9222/p2p/16Uiu2HAmGJAiUX6HLSo4nLh8T984qxzokwL23cVsYuNZy2SrK7C6,/ip4/34.83.214.149/tcp/9222/p2p/16Uiu2HAmAko2Kr3eAjM7tnshtEhYrxQYfKUvN2kwiygeFoBAoi8S,/ip4/34.169.5.52/tcp/9222/p2p/16Uiu2HAmKc6YKHzYgsjBDaj36uAufxpgZFgrzDqVBt6zTPwdhhJD \
--altda.enabled=true \
--altda.da-server=http://localhost:4242 \
--altda.da-service=true \
--altda.verify-on-read=false
```

You can check the logs running `docker logs -f op-node`. If you start op-node before op-geth, it will shut down after a few seconds if it cannot connect to its corresponding op-geth instance. This is normal behavior. It will run successfully once op-geth is running and it can connect to it.


#### Supporting historical execution

The Celo L2 node alone cannot serve RPC requests requiring historical state or execution from before the migration. To serve such requests, you must configure a HistoricalRPCService URL pointing to a Celo L1 node. The Celo L2 node will then proxy requests for historical state / execution out to the Celo L1 node.

Expand Down

0 comments on commit 225a9e9

Please sign in to comment.