Skip to content

Latest commit

 

History

History
186 lines (120 loc) · 8.05 KB

1-dorado-migration-v0.10.2.md

File metadata and controls

186 lines (120 loc) · 8.05 KB
Raw

Dorado migration

This document describe the steps needed in order to upgrade a validator node running on fetchd v0.9.x to the fetchd v0.10.2 version. It is advised to read the whole document before starting applying this procedure, in order to avoid your validator node being unable to restart. In case anything is unclear or if you have any questions, feel free to reach us out or ask in the Discord #mainnet channel.

This upgrade mostly aim on upgrading the network genesis to:

  • support cosmos-sdk v0.45.1
  • support ibc-go v2.2.0
  • support wasmd v0.24
  • mint a new nanonomx token
  • mint a new ulrn token

Before starting

Get familiar with the new version

This time, the Dorado upgrade doesn't introduce many changes on fetchd itself. The main breaking change is comming from ibc-go new version and the new cosmos-sdk version. You can test this new version by following the install instructions for the fetchd v0.10.2 version and interact with our Dorado testnet (get the connections settings on our networks documentation page).

Locate your files

Before starting, make sure to locate the following files on your validator host:

  • FETCHD_HOME: folder containing all of the validator data. By default, the fetchd command will automatically create it under ~/.fetchd, or in the location pointed by the --home flag on its first invocation if it doesn't exist.
  • FETCHD_HOME/config/node_key.json: this file hold your node private key.
  • FETCHD_HOME/config/priv_validator_key.json: this file hold your validator private key.
  • FETCHD_HOME/config/config.toml: this is your node configuration
  • FETCHD_HOME/config/app.toml: this configuration allows to tune the cosmos application, such as enabled apis, telemetry, pruning...
  • FETCHD_HOME/config/client.toml: this file stores the client config, equivalent to the legacy fetchcli config which is now replaced by fetchd config.

If needed, clone this repo and use the ./scripts/locate_home.sh script to help find the right folder, giving it your validator operator address and a search path:

./scripts/locate_home.sh fetchvaloper1fvcepqdw4lcc4s0gmxxfhkptyasfceg69x9gsc /home/

(if no match, rerun it with sudo to allow it traversing directories owned by other users)

For the rest of the document, FETCHD_HOME is assumed to be at the default ~/.fetchd/ location. If your installation is different, replace ~/.fetchd/ path in all the commands with your actual FETCHD_HOME path.

Join discord channel

We'll use a dedicated discord channel (#mainnet) for the migration, where we'll share important information such as the block height we'll stop at, or validating sha256sum of the genesis file at each steps. Make sure to join it! This repository will also get updated on the way with all the necessary informations to validate a node migration.

Stop your validator

First, make sure your node have reached at least the 5300200 block height we will use to export the network state and restart from. You can configure your node in advance to stop at this height by setting the halt-height parameter in the app.toml file and restarting your node. Also ensure that no process managers (such as systemd) will attempt to restart it.

The exact procedure to stop your node depends on how you configured it so we can't really give a generic way here.

Also double check it's properly stopped to avoid file corruption in the next steps.

Backup FETCHD_HOME

Before making any changes, it's prefered to create a backup copy of your current FETCHD_HOME directory.

The following command can be used, assuming your FETCHD_HOME is ~/.fetchd/:

cp -R ~/.fetchd/ ~/.fetchd_old/

This would allow to revert back to your starting state in case something goes wrong on the way.

Export network state

fetchd --home ~/.fetchd/ export --height 5300200 > genesis_export_5300200.json

Generate a hash of this file and validate it with others:

sha256sum genesis_export_5300200.json

Expected hash e0df0ea9b073828047afef06681b4f3f2d7f8857c71b478b2a0ecf1684f77736 File available at ./data/genesis_export_5300200.json

When your genesis hash matches the expected one, it's now time to update fetchd to the latest version.

Reset the fetchd database

Now we have exported the state properly, we can drop the fetchd database:

Double check you have proper backups of your FETCHD_HOME, and that the exported genesis is the correct one. fetchd export won't have anything to export anymore after this step!

fetchd --home ~/.fetchd/ unsafe-reset-all

Install new fetchd version

You may already have the fetchd repository on your machine from the previous installation. If not, you can:

git clone --branch v0.10.2 https://github.com/fetchai/fetchd.git fetchd_0.10.2
cd fetchd_0.10.2

If you already have an existing clone, place yourself in and:

git fetch
git checkout v0.10.2

Now you can install the new fetchd version:

make install

# and verify you now have the correct version:
fetchd -h
# must print fetchd help message

fetchd version
# must print v0.10.2

Migrate genesis

Now we have the new version of fetchd, we're ready to upgrade the genesis file to Dorado:

fetchd --home ~/.fetchd/ dorado-migrate \
    --genesis-time "2022-04-05T16:00:00Z" \
    genesis_export_5300200.json > genesis_migrated_5300200.json

We're setting here the time when the network will restart. The new chainID, initial block number and other important parameters are automatically set by the dorado-migrate command via their default values.

Again, we'll hash the created genesis and ensure it matches the expected hash with other people:

sha256sum genesis_migrated_5300200.json

Expected hash fe5fe42dc375ae33268c88cb03fe0c030c96f08ad96ad21921c58fd02fe711c9 File available at ./data/genesis_migrated_5300200.json

Copy genesis to your node home

Run the following:

cp genesis_migrated_5300200.json ~/.fetchd/config/genesis.json

Restart your node

You're now ready to restart your fully migrated node!

Caution!

  • To avoid errors here, make sure to wait for a confirmation on Discord from me (@daeMOn#5105) that the seeds are up and running. Otherwise, unexpected startup errors might occur.

  • If you have set a halt-height in your FETCHD_HOME/config/app.toml earlier to stop your node, remember to retore it to halt-height = 0 to allow your node to start!

Run:

fetchd --home ~/.fetchd/ start --p2p.seeds 17693da418c15c95d629994a320e2c4f51a8069b@connect-fetchhub4.m-v2-london-c.fetch-ai.com:36456,a575c681c2861fe945f77cb3aba0357da294f1f2@connect-fetchhub4.m-v2-london-c.fetch-ai.com:36457,d7cda986c9f59ab9e05058a803c3d0300d15d8da@connect-fetchhub4.m-v2-london-c.fetch-ai.com:36458

If you did not configure yet a min-gas-prices in your app.toml configuration, you will see the following warning message on startup:

4:50PM ERR WARNING: The minimum-gas-prices config in app.toml is set to the empty string. This defaults to 0 in the current version, but will error in the next version (SDK v0.45). Please explicitly put the desired minimum-gas-prices in your app.toml.

This is fine and can be ignored, or you're free to set a min-gas-prices to a value of your choice for now. We'll reach out to you in a near future to discuss about the min-gas-prices setting.

If you have errors at launch, first try to fetchd --home ~/.fetchd/ unsafe-reset-all first and restart. If no changes, reach out on Discord for help!

After starting, some messages will be printed in the console, and no activity will happen until 2/3 of the voting power come back online. When enough validators are online, you should see some activity in logs with messages like:

8:40AM INF executed block height=XYZ module=state num_invalid_txs=0 num_valid_txs=0

Meaning we're back producing blocks on the new network.