Extend the twin test to use snapshot file to load the DB. Not a full zipped archive of a DB.

[saki-osive]

Is your feature request related to a problem? Please describe. Currently the twin tests use the full zipped archive of a DB.

Describe the solution you'd like Extend the twin test to use snapshot file to load the DB

• Download a snapshot archive & unzip
• Turn archive snapshot into genesis.blob (bytes for initial state of blockchain).
• Start twin from a saved genesis.blob which came from the snapshot

Ideally the CLI interface for the twin tests has three db options a) use a full directory --db-path b) use a snapshot --snapshot-path c) (nice to have) use a genesis.blob --genesis-blob, which would have been generated in a previous test run. Speeds up testing

Twin tests location: https://github.com/0LNetworkCommunity/libra-framework/blob/main/tools/rescue/src/twin.rs

Storage: read snapshot https://github.com/0LNetworkCommunity/libra-framework/blob/main/tools/storage/src/read_snapshot.rs

It might might be possible to import read_snapshot into twin.rs

• TASK ONE: try to make the twin CLI accept a --snapshot-path file. Likely integrating read_snapshot.rs
• TASK TWO: Add a standard test: can we upgrade a Twin chain with a new "framework upgrade" procedure.
• TASK THREE: Set up github actions to have a new workflow which runs the Twin upgradeability tests. Using the prior night's snapshot in epoch_archive

Here is a test using the twin with a local copy db https://github.com/0LNetworkCommunity/libra-framework/blob/0c279849dd48081bd738a1388ad73f68db742439/tools/rescue/tests/upgrade_multi_action.rs

Repo of snapshots: https://github.com/0LNetworkCommunity/epoch-archive-mainnet Alternative to TASK ONE: take the libra binary, and start in fullnode mode, to populate a ./db/ which then can be used for twin

[0o-de-lally]

1. Read snapshot
2. Test for snapshot (not passing)
3. TODO: Add test for upgradeability
4. TODO: (nice to have) use a genesis.blob to speed up.

[0o-de-lally]

@saki-osive Try to confirm that the twin tool is creating the relevant directories for swarm. This should be one of the check a meta test. (There is a swam method that reveals the directory of each node, note that it may be a temp folder).

[saki-osive]

1. Read snapshot
2. Test for snapshot (not passing)
3. TODO: Add test for upgradeability
4. TODO: (nice to have) use a genesis.blob to speed up.

1. Read snapshot

• Integrated read_snaphost in the twin
• Added check to use snapshot file to load the DB if provided
• Updated the twin cli to accept the snapshot file