A Kurtosis package that deploys a private, portable, and modular Polygon CDK devnet over Docker or Kubernetes.
Specifically, this package will deploy:
- A local L1 chain, fully customizable with multi-client support, using the ethereum-package.
- A local L2 chain, using the Polygon Chain Development Kit (CDK), with customizable components such as sequencer, sequence sender, aggregator, rpc, prover, dac, etc. It will first deploy the Polygon zkEVM smart contracts on the L1 chain before deploying the different components.
- The zkEVM bridge infrastructure to facilitate asset bridging between the L1 and L2 chains, and vice-versa.
- The Agglayer, an in-development interoperability protocol, that allows for trustless cross-chain token transfers and message-passing, as well as more complex operations between L2 chains, secured by zk proofs.
- Additional services such as transaction spammer, monitoring tools, permissionless nodes etc.
🚨 This package is currently designed as a development tool for testing configurations and scenarios within the Polygon CDK stack. It is not recommended for long-running or production environments such as testnets or mainnet. If you need help, you can reach out to the Polygon team or talk to an Implementation Partner (IP).
The package is flexible and supports various configurations for deploying and testing the Polygon CDK stack.
You can take a look at this table to see which versions of the CDK are meant to work together, broken by fork identifier.
The table provided illustrates the different combinations of sequencers and sequence sender/aggregator components that can be used, along with their current support status in Kurtosis.
The team is actively working on enabling the use cases that are currently not possible.
Stack | Sequencer | Sequence Sender / Aggregator | Supported by Kurtosis? |
---|---|---|---|
New CDK stack | cdk-erigon | cdk-node | ✅ |
New sequencer with new zkevm stack | cdk-erigon | zkevm-sequence-sender + zkevm-aggregator | ❌ (WIP) - Check the kurtosis-cdk-erigon package |
New sequencer with legacy zkevm stack | cdk-erigon | zkevm-node | ❌ (WIP) |
Legacy sequencer with new cdk stack | zkevm-node | cdk-node | ❌ (WIP) |
Legacy sequencer with new zkevm stack | zkevm-node | zkevm-sequence-sender + zkevm-aggregator | ❌ (WIP) |
Legacy zkevm stack | zkevm-node | zkevm-node | ✅ |
To understand how to configure Kurtosis for these use cases, refer to the documentation and review the test files located in the .github/tests/
directory.
To begin, you will need to install Docker (>= v4.27.0 for Mac users) and Kurtosis.
- If you notice some services, such as the
zkevm-stateless-executor
orzkevm-prover
, consistently having the status ofSTOPPED
, try increasing the Docker memory allocation.
If you intend to interact with and debug the stack, you may also want to consider a few additional optional tools such as:
Once that is good and installed on your system, you can run the following command to deploy the complete CDK stack locally. This process typically takes around eight to ten minutes.
kurtosis run --enclave cdk github.com/0xPolygon/kurtosis-cdk
The default deployment includes cdk-erigon as the sequencer, and cdk-node functioning as the sequence sender and aggregator. You can verify the default versions of these components and the default fork ID by reviewing input_parser.star. You can check the default versions of the deployed components and the default fork ID by looking at input_parser.star.
To make customizations to the CDK environment, clone this repo, make any desired configuration changes, and then run:
# Delete all stop and clean all currently running enclaves
kurtosis clean --all
# Run this command from the root of the repository to start the network
kurtosis run --enclave cdk .
Let's do a simple L2 RPC test call.
First, you will need to figure out which port Kurtosis is using for the RPC. You can get a general feel for the entire network layout by running the following command:
kurtosis enclave inspect cdk
That output, while quite useful, might also be a little overwhelming. Let's store the RPC URL in an environment variable.
You may need to adjust the various commands slightly if you deployed the legacy zkevm-node as the sequencer. You should target the
zkevm-node-rpc-001
service instead ofcdk-erigon-rpc-001
.
export ETH_RPC_URL="$(kurtosis port print cdk cdk-erigon-rpc-001 rpc)"
That is the same environment variable that cast
uses, so you should now be able to run this command. Note that the steps below will assume you have the Foundry toolchain installed.
cast block-number
By default, the CDK is configured in test
mode, which means there is some pre-funded value in the admin account with address 0xE34aaF64b29273B7D567FCFc40544c014EEe9970
.
cast balance --ether 0xE34aaF64b29273B7D567FCFc40544c014EEe9970
Okay, let’s send some transactions...
private_key="0x12d7de8621a77640c9241b2595ba78ce443d05e94090365ab3bb5e19df82c625"
cast send --legacy --private-key "$private_key" --value 0.01ether 0x0000000000000000000000000000000000000000
Okay, let’s send even more transactions... Note that this step will assume you have polygon-cli installed.
polycli loadtest --rpc-url "$ETH_RPC_URL" --legacy --private-key "$private_key" --verbosity 700 --requests 50000 --rate-limit 50 --concurrency 5 --mode t
polycli loadtest --rpc-url "$ETH_RPC_URL" --legacy --private-key "$private_key" --verbosity 700 --requests 500 --rate-limit 10 --mode 2
polycli loadtest --rpc-url "$ETH_RPC_URL" --legacy --private-key "$private_key" --verbosity 700 --requests 500 --rate-limit 3 --mode uniswapv3
Pretty often, you will want to check the output from the service. Here is how you can grab some logs:
kurtosis service logs cdk agglayer --follow
In other cases, if you see an error, you might want to get a shell in the service to be able to poke around.
kurtosis service shell cdk contracts-001
jq . /opt/zkevm/combined.json
One of the most common ways to check the status of the system is to make sure that batches are going through the normal progression of trusted, virtual, and verified:
cast rpc zkevm_batchNumber
cast rpc zkevm_virtualBatchNumber
cast rpc zkevm_verifiedBatchNumber
If the number of verified batches is increasing, then it means the system works properly.
To access the zkevm-bridge
user interface, open this URL in your web browser.
open "$(kurtosis port print cdk zkevm-bridge-proxy-001 web-ui)"
When everything is done, you might want to clean up with this command which stops the local devnet and deletes it.
kurtosis clean --all
For more information about the CDK stack, visit the Polygon Knowledge Layer.
This section features documentation specifically designed for advanced users, outlining complex operations and techniques.
- How to use CDK ACL.
- How to deploy additional services alongside the CDK stack, such as transaction spammer, monitoring tools, permissionless nodes etc.
- How to attach multiple CDK chains to the AggLayer.
- How to use the different data availability modes.
- How to deploy the stack to an external L1 such as Sepolia.
- How to deploy contracts with the deterministic deployment proxy.
- How to edit the zkevm contracts.
- How to perform an environment migration with clean copies of the databases.
- How to iterate and debug quickly with Kurtosis.
- How to use zkevm contracts fork 12.
- How to integrate a third-party data availability committee (DAC).
- How to migrate from fork 7 to fork 9.
- How to upgrade forks for isolated CDK chains.
- How to use a native token.
- How to play with the network to introduce latencies.
- How to set up a permissionless zkevm node.
- How to resequence batches with the cdk-erigon sequencer.
- How to run a debugger.
- How to assign static ports to Kurtosis services.
- How to work with the timelock.
- How to trigger a reorg.
- How to perform a trustless recovery the DAC and L1.
Click to expand
- Deploy the package without cloning the repository.
kurtosis run --enclave cdk github.com/0xPolygon/kurtosis-cdk
kurtosis run --enclave cdk github.com/0xPolygon/kurtosis-cdk@main
kurtosis run --enclave cdk github.com/0xPolygon/kurtosis-cdk@v0.2.15
- Deploy the package with the default parameters.
kurtosis run --enclave cdk .
- Deploy with the default parameters and specify on-the-fly custom arguments.
kurtosis run --enclave cdk . '{"deployment_stages": {"deploy_l1": false}}'
- Deploy with a configuration file.
Check the tests folder for sample configuration files.
kurtosis run --enclave cdk --args-file params.yml .
- Deploy with a configuration file and specify on-the-fly custom arguments.
Note: In this specific case, on-the-fly custom arguments take precedence over defaults and config file arguments.
kurtosis run --enclave cdk --args-file params.yml . '{"args": {"agglayer_image": "ghcr.io/agglayer/agglayer:latest"}}'
Click to expand
By default your Kurtosis cluster should be docker
. You can check this using the following command:
kurtosis cluster get
You can also list the available clusters.
kurtosis cluster ls
If you take a look at your Kurtosis configuration, it should be similar to this:
config-version: 2
should-send-metrics: true
kurtosis-clusters:
docker:
type: "docker"
Let's say you've deployed a local minikube cluster. It would work the same for any type of Kubernetes cluster.
Edit the Kurtosis configuration file.
vi "$(kurtosis config path)"
Under kurtosis-clusters
, you should add another entry for your local Kubernetes cluster.
kurtosis-clusters:
minikube: # give it the name you want
type: "kubernetes"
config:
kubernetes-cluster-name: "local-01" # should be the same as your cluster name
storage-class: "standard"
enclave-size-in-megabytes: 10
Then point Kurtosis to the local Kubernetes cluster.
kurtosis cluster set minikube
Deploy the package to Kubernetes.
kurtosis run --enclave cdk .
If you want to revert back to Docker, simply use:
kurtosis cluster set docker
Click to expand
Occasionally, Kurtosis deployments may run indefinitely. Typically, deployments should complete within 10 to 15 minutes. If you experience longer deployment times or if it seems stuck, check the Docker engine's memory limit and set it to 16GB if possible. If this does not resolve the issue, please refer to the troubleshooting steps provided.
-
Make sure the issue is related to Kurtosis itself. If you made any changes to the package, most common issues are misconfigurations of services, file artefacts, ports, etc.
-
Remove the Kurtosis enclaves.
Note: By specifying the --all
flag, Kurtosis will also remove running enclaves.
kurtosis clean --all
- Restart the Kurtosis engine.
kurtosis engine restart
- Restart the Docker daemon.
Click to expand
Kurtosis is just a thin wrapper on top of Docker so you can use all the docker
commands you want.
On top of that, here are some useful commands.
- View the state of the enclave (services and endpoints).
kurtosis enclave inspect cdk
- Follow the logs of a service.
Note: If you want to see all the logs of a service, you can specify the --all
flag.
kurtosis service logs cdk cdk-erigon-sequencer-001 --follow
- Execute a command inside a service.
kurtosis service exec cdk contracts-001 'cat /opt/zkevm/combined.json' | tail -n +2 | jq
- Get a shell inside a service.
kurtosis service shell cdk cdk-erigon-sequencer-001
- Stop or start a service.
kurtosis service stop cdk cdk-erigon-sequencer-001
kurtosis service start cdk cdk-erigon-sequencer-001
- Get a specific endpoint.
kurtosis port print cdk cdk-erigon-node-001 http-rpc
- Inspect a file artifact.
kurtosis files inspect cdk cdk-erigon-node-config-artifact-sequencer config.yaml
- Download a file artifact.
kurtosis files download cdk cdk-erigon-node-config-artifact
Click to expand
kurtosis lint --format .
Click to expand
Head to the Kurtosis documentation.
You can also take a look at the Starlark language specification to know if certain operations are supported.
- For technical issues, join our Discord.
- For documentation issues, raise an issue on the published live doc at our main repo.
Copyright (c) 2024 PT Services DMCC
Licensed under either:
- Apache License, Version 2.0, (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0), or
- MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)
as your option.
The SPDX license identifier for this project is MIT
OR Apache-2.0
.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.