Testing your Oracle DAO Node#

Once your node is set up and you've joined the Oracle DAO, you should test it to ensure it's able to perform its duties properly. The best way to do this is to have it build the Redstone rewards Merkle tree using Rocket Pool's treegen utility.

treegen#

treegen is a tool that can reproduce the entire rewards Merkle tree and accompanying artifacts for a previous rewards interval via your archive Execution and Consensus clients. It can also "dry run" the current interval by pretending that it ended at the latest finalized epoch (at the time of running it) and producing a partial tree from the start of the interval up to that point.

treegen can be used as a standalone binary (currently only built for Linux systems, x64 and arm64) or as a Docker container.

If you would like to download the standalone binary, you can find it in the releases here: https://github.com/rocket-pool/treegen. Usage instructions are included in the README there, but we'll cover some examples below as well.

The Docker container tag for it is rocketpool/treegen:latest.

Building a Dry-Run Tree#

For a first test, run treegen to generate a dry-run tree that calculates the tree from the start of the rewards interval to the latest (finalized) slot. We'll use the script included in the repository that leverages the Docker container to run it on the node machine itself for simplicity:

./treegen.sh -e http://localhost:8545 -b http://localhost:5052

This will test your clients' abilities to respond to queries in a timely fashion (e.g., if you are using a third-party service, this will be helpful to assess whether or not its query rate-limit is insufficient), but will not test their Archive Mode capabilities. It will produce output like the following:

If this runs without error, it will generate the rewards tree artifacts and save them as JSON files in your working directory. You are free to explore them and ensure their contents are sane, but as they are dry-run files, they aren't canonically stored anywhere for comparison.

Building a Canonical Tree from a Past Interval#

This next test is to replicate one of the complete trees from a past interval. This will require archival access on both the Execution Layer and the Consensus Layer, so it will serve as a good test of both capabilities.

As of this writing, Interval 2 is an ideal choice as it is far in the past () and involved the Smoothing Pool (which accounts for the largest computational load when calculating the rewards for the period).

Run treegen using the following command:

./treegen.sh -e http://<your archive EC url> -b http://localhost:5052 -i 2

Note that the Execution Client URL is different here: it must be an Archive EC as the snapshot block for Interval 2 was far in the past.

The output will look like this (truncated for previty):

The key thing to look for here is this message at the end:

If you receive this, then your watchtower can build the tree correctly.

Next Steps#

Next up, we'll cover how to monitor your node's performance.