Deployment.md

Deployment Script

Pre-requisite

You need to have an account on Aptos testnet/mainnet that

1. is sufficiently funded for package publishing and transaction execution,
2. you know the name of this account in your Aptos profile, which can be typically found in a file named .aptos/config.yaml
3. is one of the active owners of the PoolAdmin multi-sig account

In the rest of this document, we will use <deployer_profile_name> to represent the name of this account in your Aptos profile.

Publish Packages

# the following commands publishes all packages on the object account
./deploy.py testnet --deployer "<deployer_profile_name>" publish-config
./deploy.py testnet --deployer "<deployer_profile_name>" publish-acl
./deploy.py testnet --deployer "<deployer_profile_name>" publish-math
./deploy.py testnet --deployer "<deployer_profile_name>" publish-mock-underlyings
./deploy.py testnet --deployer "<deployer_profile_name>" publish-oracle
./deploy.py testnet --deployer "<deployer_profile_name>" publish-core
./deploy.py testnet --deployer "<deployer_profile_name>" publish-data

The addresses of published packages can be found at deploy-*-object.txt files, respectively. For example, the address where aave_pool is published will be stored in deploy-aave_pool-object.txt.

Run Setup Scripts

The first thing to do is to configure the Access Control List via:

./deploy.py testnet --deployer "<deployer_profile_name>" configure-acl

This will grant necessary roles to known accounts addresses and will transfer the super-admin role to the AaveACL multisig account. After this step, the deployer no longer has the super-admin role, and every role change will need to go through the AaveACL multisig account.

The rest of the configuration steps involve the PoolAdmin multisig account. Running the following commands will only submit transactions to the PoolAdmin multisig account for approval. The final execution of the proposed payload will be triggered by a second transaction that executes the approved proposal.

./deploy.py testnet --deployer "<deployer_profile_name>" setup-configure-emodes
./deploy.py testnet --deployer "<deployer_profile_name>" setup-create-reserves
./deploy.py testnet --deployer "<deployer_profile_name>" setup-configure-reserves
./deploy.py testnet --deployer "<deployer_profile_name>" setup-configure-interest-rates
./deploy.py testnet --deployer "<deployer_profile_name>" setup-configure-price-feeds

Transfer Package Ownerships

Once deployed, the ownership of the objects that host the published packages can be transferred to corresponding multisig accounts using the following commands:

./deploy.py testnet --deployer "<deployer_profile_name>" change-owner-config
./deploy.py testnet --deployer "<deployer_profile_name>" change-owner-acl
./deploy.py testnet --deployer "<deployer_profile_name>" change-owner-math
./deploy.py testnet --deployer "<deployer_profile_name>" change-owner-mock-underlyings
./deploy.py testnet --deployer "<deployer_profile_name>" change-owner-oracle
./deploy.py testnet --deployer "<deployer_profile_name>" change-owner-core
./deploy.py testnet --deployer "<deployer_profile_name>" change-owner-data

Upgrade Packages

If package upgrading is needed, use one of the following command to upgrade the package that must be already published on the object account (with address stored at the deploy-*-object.txt). This will use the multisig upgrading procedure as the packages are now assumed to have been transferred to the corresponding multisig accounts (and hence, are no longer owned by the deployer).

./deploy.py testnet --deployer "<deployer_profile_name>" upgrade-config
./deploy.py testnet --deployer "<deployer_profile_name>" upgrade-acl
./deploy.py testnet --deployer "<deployer_profile_name>" upgrade-math
./deploy.py testnet --deployer "<deployer_profile_name>" upgrade-mock-underlyings
./deploy.py testnet --deployer "<deployer_profile_name>" upgrade-oracle
./deploy.py testnet --deployer "<deployer_profile_name>" upgrade-core
./deploy.py testnet --deployer "<deployer_profile_name>" upgrade-data

Local Testnet (localnet) Simulation

You can also simulate the whole deployment process on a locally spawn testnet (localnet), using the following command:

./deploy.py localnet

You will be greeted with

Press Enter when localnet is deployed (i.e., until "Setup is complete"):

So wait around 30 seconds, and hit "Enter" when you see

Setup is complete, you can now use the localnet!

It will prompt you with either creating a new <deployer_profile_name> profile or reusing an existing one.

After that, package publishing and the setup steps will be executed by themselves. Setup also goes through a multisig account on localnet, but no additional approval is needed and the proposal execution is automated as well.

GHO reserve deployment

If the protocol is already deployed, one has to first run to re-initialize the initial deployer config for the gho configuration to be made available for the setup-gho-reserve method

./deploy.py testnet --deployer "<deployer_profile_name>" admin-reset-data

Then, adding the gho reserve to the protocol is simply a matter of running:

./deploy.py testnet --deployer "<deployer_profile_name>" setup-gho-reserve

Finally, the direct minter address (whether it is an object or account) has to be given the risk admin role via:

./deploy.py testnet --deployer "<deployer_profile_name>" --gho-direct-minter "<gho_direct_minter_address>" add-risk-admin