This is a meta package of cosmos-sdk-based docker images and configuration meant to make deploying onto Akash easy and standardized across cosmos.
The goal is to have a simple way to launch any cosmos chain, with a variety of different bootstrapping options
- "normal" boostrap - using
fastsync. - Hand-made snapshots a la cosmos-snapshots
- The new
state-syncmechanism.
Configuration is achieved using environment variables, with shortcuts available for common setups. Every aspect of the node configuration can be achieved in this way.
Additionally the node private keys can be backed up and restored from any S3 compatible storage provider, such as Sia or Storj, via Filebase.
The available docker images can be found here. They are
tagged with the form $COSMOS_OMNIBUS_VERSION-$PROJECT-$PROJECT_VERSION.
| Project | Version | Image | |
|---|---|---|---|
| akash | v0.12.1 |
ghcr.io/ovrclk/cosmos-omnibus:v0.0.5-akash-v0.12.1 |
Example |
| sentinelhub | v0.7.0 |
ghcr.io/ovrclk/cosmos-omnibus:v0.0.5-sentinelhub-v0.7.0 |
Example |
| gaia | v4.2.1 |
ghcr.io/ovrclk/cosmos-omnibus:v0.0.5-gaia-v4.2.1 |
Example |
| kava | v0.14.2 |
ghcr.io/ovrclk/cosmos-omnibus:v0.0.5-kava-v0.14.2 |
Example |
| osmosis | v3.1.0 |
ghcr.io/ovrclk/cosmos-omnibus:v0.0.5-osmosis-v3.1.0 |
Example |
| persistence | v0.1.3 |
ghcr.io/ovrclk/cosmos-omnibus:v0.0.5-persistence-v0.1.3 |
Example |
| juno | lucina |
ghcr.io/ovrclk/cosmos-omnibus:v0.0.5-juno-lucina |
Example |
Cosmos blockchains can be configured entirely using environment variables instead of the config files. Every chain has it's own prefix, but the format of the configuration is the same.
For example to configure the seeds option in the p2p section of config.toml, for the Akash blockchain:
AKASH_P2P_SEEDS=id@node:26656
The namespace for each of the supported chains in the cosmos omnibus can be found in the docker-compose.yml files in each project directory.
The omnibus images allow some specific variables and shortcuts to configure extra functionality.
Chain config can be sourced from a chain.json file as implemented by @sikkatech, or from a metadata repository such as Akash's ovrclk/net.
| Variable | Description | Default | Examples |
|---|---|---|---|
CHAIN_URL |
URL to a chain.json file detailing the chain meta |
https://github.com/sikkatech/tm-networks/blob/master/akashnet-2/chain.json |
|
METADATA_URL |
URL to a net repo in the same form as Akash |
https://github.com/ovrclk/net/tree/master/mainnet |
|
CHAIN_ID |
The cosmos chain ID | akashnet-2 |
|
GENESIS_URL |
URL to the genesis file in .gz or .zip format. Can be set by CHAIN_URL or METADATA_URL |
https://raw.githubusercontent.com/ovrclk/net/master/mainnet/genesis.json |
|
VALIDATE_GENESIS |
Set to 0 to disable validation of genesis file | 1 |
0 |
See Cosmos docs for more information. Note this can be sourced from CHAIN_URL or METADATA_URL automatically.
| Variable | Description | Default | Examples |
|---|---|---|---|
P2P_SEEDS |
Seed nodes. Can be set by CHAIN_URL or GENESIS_URL | id@node:26656 |
|
P2P_PERSISTENT_PEERS |
Persistent peers. Can be set by CHAIN_URL or GENESIS_URL | id@node:26656 |
On boot, the container can restore a private key from any S3 storage provider. If this is configured and the key doesn't exist in S3 yet, it will be uploaded from the node.
This allows for a persistent node ID and validator key on Akash's ephemeral storage.
Filebase is a great way to use S3 with decentralised hosting such as Sia.
The node_key.json and priv_validator_key.json are both backed up, and can be encrypted with a password.
| Variable | Description | Default | Examples |
|---|---|---|---|
S3_KEY |
S3 access key | ||
S3_SECRET |
S3 secret key | ||
S3_HOST |
The S3 API host | https://s3.filebase.com |
s3.us-east-1.amazonaws.com |
KEY_PATH |
Bucket and directory to backup/restore to | bucket/nodes/node_1 |
|
KEY_PASSWORD |
An optional password to encrypt your private keys. Shouldn't be optional |
Some shortcuts for enabling statesync. Statesync requires 2x nodes with snapshots enabled.
| Variable | Description | Default | Examples |
|---|---|---|---|
STATESYNC_SNAPSHOT_INTERVAL |
Take a snapshot to provide statesync every X blocks | 500 |
|
STATESYNC_ENABLE |
Enabling statesyncing from a node. Default true if STATESYNC_RPC_SERVERS is set |
||
STATESYNC_RPC_SERVERS |
Comma separated list of RPC nodes with snapshots enabled | ip:26657,ip2:26657 |
|
TRUSTED_NODE |
A trusted node to obtain trust height and hash from. Defaults to the first STATESYNC_RPC_SERVERS if set |
ip:26657 |
|
STATESYNC_TRUST_PERIOD |
Trust period for the statesync snapshot | 168h0m0s |
|
STATESYNC_TRUST_HEIGHT |
Obtained from TRUSTED_NODE or first STATESYNC_RPC_SERVERS | ||
STATESYNC_TRUST_HASH |
Obtained from TRUSTED_NODE or first STATESYNC_RPC_SERVERS |
The node data directory can be restored from a .tar or .tar.gz file stored on a public URL.
This can be from a specific URL, or from a base URL and a file matching a given pattern.
| Variable | Description | Default | Examples |
|---|---|---|---|
BOOTSTRAP |
Whether to bootstrap the node from the snapshot URL | 1 |
0 |
SNAPSHOT_URL |
A URL to a .tar or .tar.gz file |
http://135.181.60.250/akash/akashnet-2_2021-06-16.tar |
|
SNAPSHOT_FORMAT |
The format of the snapshot file | tar.gz |
tar |
SNAPSHOT_BASE_URL |
A base URL to a directory containing backup files | http://135.181.60.250/akash |
|
SNAPSHOT_PATTERN |
The pattern of the file in the BASE_URL |
$CHAIN_ID.*$SNAPSHOT_FORMAT |
foobar.*tar.gz |
See Cosmos docs for more information
| Variable | Description | Default | Examples |
|---|---|---|---|
FASTSYNC_VERSION |
The fastsync version | v2 |
|
MINIMUM_GAS_PRICES |
Minimum gas prices | 0.025uakt |
|
PRUNING |
How much of the chain to prune | nothing |
|
DEBUG |
Set to 1 to output all environment variables on boot |
1 |
See the deploy.yml example file in each chain directory which details the minimum configuration required. Use the above configuration options to add functionality.
See the docker-compose.yml example file to run each node locally using docker-compose up
See the _examples directory for some common setups, including:
- Backup node data to S3 on schedule
- More chains..
Adding a new chain is easy:
- Ideally setup a
chain.jsonornetrepository to provide a single source of truth for setup info - Add a project directory based on the existing projects
- The github workflow to create an image for your chain
Submit a PR or an issue if you want to see any specific chains.