Skip to content

Latest commit

 

History

History
286 lines (216 loc) · 12.4 KB

README.md

File metadata and controls

286 lines (216 loc) · 12.4 KB

Try-runtime

Substrate's programmatic testing framework.

As the name suggests, try-runtime is a detailed testing framework that gives you a lot of control over what is being executed in which environment. It is recommended that user's first familiarize themselves with substrate in depth, particularly the execution model. It is critical to deeply understand how the wasm/client/runtime interactions, and the runtime apis work in the substrate runtime, before commencing to working with try-runtime.

Resources

Some resources about the above:

  1. https://www.crowdcast.io/e/substrate-seminar/41
  2. https://docs.substrate.io/fundamentals/runtime-development/
  3. https://www.youtube.com/watch?v=a_u3KMG-n-I

Background Knowledge

The basis of all try-runtime commands is the same: connect to a live node, scrape its state and put it inside a TestExternalities, then call into a specific runtime-api using the given state and some runtime.

Alternatively, the state could come from a snapshot file.

All of the variables in the above statement are made italic. Let's look at each of them:

  1. State is the key-value pairs of data that comprise the canonical information that any blockchain is keeping. A state can be full (all key-value pairs), or be partial (only pairs related to some pallets/prefixes). Moreover, some keys are special and are not related to specific pallets, known as well_known_keys in substrate. The most important of these is the :CODE: key, which contains the code used for execution, when wasm execution is chosen.

  2. A runtime-api call is a call into a function defined in the runtime, on top of a given state. Each subcommand of try-runtime utilizes a specific runtime-api.

  3. Finally, the runtime is the actual code that is used to execute the aforementioned runtime-api. Everything in this crate assumes wasm execution, which means the runtime that you use is the one stored onchain, namely under the :CODE: key.

To recap, a typical try-runtime command does the following:

  1. Download the state of a live chain, and write to an externalities.
  2. Overwrite the :CODE: with a given wasm blob
  3. Test some functionality via calling a runtime-api.

Installation

# Install latest version (recommended for local development)
cargo install --git https://github.com/paritytech/try-runtime-cli --locked
# Install a specific version (recommended for tools like CI)
cargo install --git https://github.com/paritytech/try-runtime-cli --tag vX.Y.Z --locked
try-runtime --help
try-runtime on-runtime-upgrade --help

Usage

To use any of the provided commands, SharedParams must be provided. The most important of which being SharedParams::runtime, which specifies which runtime to use. Furthermore, SharedParams::overwrite_state_version can be used to alter the state-version (see https://forum.polkadot.network/t/state-trie-migration/852 for more info).

Then, the specific command has to be specified. See Action for more information about each command's specific customization flags, and assumptions regarding the runtime being used.

Briefly, this CLI is capable of executing:

  • Action::OnRuntimeUpgrade: execute all the OnRuntimeUpgrade hooks.
  • Action::ExecuteBlock: re-execute the given block.
  • Action::FastForward: execute OnRuntimeUpgrade hooks, then fast-forward the chain a given number of blocks while checking try-state invarients.
  • Action::OffchainWorker: re-execute the given block's offchain worker code path.
  • Action::FollowChain: continuously execute the blocks of a remote chain on top of a given runtime.
  • Action::CreateSnapshot: Create a snapshot file from a remote node.

Finally, to make sure there are no errors regarding this, always run any try-runtime command with executor=trace logging targets, which will specify which runtime is being used per api call. Moreover, remote-ext, try-runtime and runtime logs targets will also be useful.

Spec name check

A common pitfall is that you might be running some test on top of the state of chain x, with the runtime of chain y. To avoid this all commands do a spec-name check before executing anything by default. This will check the, if any alterations are being made to the :CODE:, then the spec names match. The spec versions are warned, but are not mandated to match.

If anything, in most cases, we expect spec-versions to NOT match, because try-runtime is all about testing unreleased runtimes.

Note on signature and state-root checks

All of the commands calling into TryRuntime_execute_block (Action::ExecuteBlock and Action::FollowChain) disable both state root and signature checks. This is because in 99% of the cases, the runtime that is being tested is different from the one that is stored in the canonical chain state. This implies:

  1. the state root will NEVER match, because :CODE: is different between the two.
  2. replaying all transactions will fail, because the spec-version is part of the transaction signature.

Best Practices

Try-runtime is all about battle-testing unreleased runtimes. The following list of suggestions help developers maximize their testing coverage and make the best use of try-runtime features.

Testing Runtime Upgrades

One of the most powerful abilities of try-runtime is using the OnRuntimeUpgrade::pre_upgrade and OnRuntimeUpgrade::post_upgrade hooks to test runtime upgrades implemented with OnRuntimeUpgrade. OnRuntimeUpgrade can be implemented inside the pallet, or standalone in a runtime to define a migration to execute next runtime upgrade. In both cases, these methods can be added:

#[cfg(feature = "try-runtime")]
fn pre_upgrade() -> Result<Vec<u8>, TryRuntimeError> {}

#[cfg(feature = "try-runtime")]
fn post_upgrade(state: Vec<u8>) -> Result<(), TryRuntimeError> {}

(The pallet macro syntax will support this simply as a part of #[pallet::hooks]).

These hooks will be called when you execute the Action::OnRuntimeUpgrade command, before and after the migration. OnRuntimeUpgrade::pre_upgrade returns a [Vec<u8>] that can contain arbitrary encoded data (usually some pre-upgrade state) which will be passed to OnRuntimeUpgrade::pre_upgrade after upgrading and used for post checking.

VersionedMigration

It is strongly suggested to use VersionedMigration when writing custom migrations for pallets.

State Consistency

Similarly, each pallet can expose a function in #[pallet::hooks] section as follows:

#[cfg(feature = "try-runtime")]
fn try_state(_: BlockNumber) -> Result<(), TryRuntimeError> {}

which is called on numerous code paths in the try-runtime tool. These checks should ensure that the state of the pallet is consistent and correct. See TryState for more info.

Logging

It is super helpful to make sure your migration code uses logging (always with a runtime log target prefix, e.g. runtime::balance) and state exactly at which stage it is, and what it is doing.

Examples

For the following examples, we assume the existence of the following:

  1. a substrate node compiled with --features try-runtime, called substrate. This will be the running node that you connect to, and provide a wasm blob that has try-runtime functionality enabled.
  2. the try-runtime CLI binary on your path.
# this is like your running deployed node.
cargo build --features try-runtime --release && cp target/release/substrate .

The above example is with substrate's kitchensink-runtime, but is applicable to any substrate-based chain.

  • Run the migrations of a given runtime on top of a live state.
# assuming there's `./substrate --dev --tmp --ws-port 9999` or similar running.
try-runtime \
    --runtime /path-to-substrate/target/release/wbuild/my-runtime.wasm \
    on-runtime-upgrade \
    live --uri ws://localhost:9999
  • Same as the previous example, but run it at specific block number's state and using the live polkadot network. This means that this block hash's state should not yet have been pruned by the node running at rpc.polkadot.io.
try-runtime \
    --runtime /path-to-polkadot-runtimes/target/release/wbuild/polkadot-runtime/polkadot-runtime.wasm \
    on-runtime-upgrade \
    live --uri wss://rpc.polkadot.io:443 \
    # replace with your desired block hash!
    --at 0xa1b16c1efd889a9f17375ec4dd5c1b4351a2be17fa069564fced10d23b9b3836
  • Now, let's use a snapshot file. First, we create the snapshot:
try-runtime --runtime existing create-snapshot --uri ws://localhost:9999 my-snapshot.snap
2022-12-13 10:28:17.516  INFO                 main remote-ext: since no at is provided, setting it to latest finalized head, 0xe7d0b614dfe89af65b33577aae46a6f958c974bf52f8a5e865a0f4faeb578d22
2022-12-13 10:28:17.516  INFO                 main remote-ext: since no prefix is filtered, the data for all pallets will be downloaded
2022-12-13 10:28:17.550  INFO                 main remote-ext: writing snapshot of 1611464 bytes to "[email protected]"
2022-12-13 10:28:17.551  INFO                 main remote-ext: initialized state externalities with storage root 0x925e4e95de4c08474fb7f976c4472fa9b8a1091619cd7820a793bf796ee6d932 and state_version V1

Note that the snapshot contains the existing runtime, which does not have the correct try-runtime feature. In the following commands, we still need to overwrite the runtime.

Then, we can use it to have the same command as before, on-runtime-upgrade

try-runtime \
    --runtime /path-to-substrate/target/release/wbuild/my-runtime.wasm \
    on-runtime-upgrade \
    snap -p my-snapshot.snap
  • Execute the latest finalized block with the given runtime.
try-runtime \
    --runtime /path-to-substrate/target/release/wbuild/my-runtime.wasm \
    execute-block live \
    --uri ws://localhost:9999

This can still be customized at a given block with --at. If you want to use a snapshot, you can still use --block-ws-uri to provide a node form which the block data can be fetched.

Moreover, this runs the TryState hooks as well. The hooks to run can be customized with the --try-state. For example:

try-runtime \
    --runtime /path-to-substrate/target/release/wbuild/my-runtime.wasm \
   execute-block \
   --try-state System,Staking \
   live \
   --uri ws://localhost:9999 \
   --pallet System Staking

Will only run the try-state of the two given pallets. When running try-state against some real chain data it can take a long time for the command to execute since it has to query all the key-value pairs. In scenarios like above where we only want to run the try-state for some specific pallets, we can use the --pallet option to specify from which pallets we want to query the state. This will greatly decrease the execution time.

See TryStateSelect for more information.

  • Follow our live chain's blocks using follow-chain, whilst running the try-state of 3 pallets in a round robin fashion
try-runtime \
    --runtime /path-to-substrate/target/release/wbuild/my-runtime.wasm \
    follow-chain \
    --uri ws://localhost:9999 \
    --try-state rr-3