You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As an Antelope Core developers, I want to be able to run an automated replay test from genesis within 1 day, to confirm that consensus is maintained and uptime reliability continues.
What
To verify that the consensus mechanism is intact, we can run a full replay from genesis and confirm that every block can be applied without error, and that we achieve the expected state after replay. It may be valuable to run multiple replay tests with different options (for example: supported CPU features, using various WASM runtimes). This would allow us to gain confidence that consensus is robust under many different environmental situations.
Why
Running a full replay from genesis manually would be a months-long endeavor. We need to be able to run a replay test within 1 day in an automated and repeatable fashion.
Considerations
These tests may be expensive to run
If developed correctly, it should be trivial to add other Antelope chains
Requirements
A developer can "hit a button" to launch a replay test for a particular commit, and see if completed within a day.
Ability to specify command line options for test (including WASM runtime)
Non-requirements (Potential Future Version of Chicken Dance)
Ability to easily run with history from other Antelope chains (Jungle, WAX, TELOS, etc)
Ability to run on different hardware to simulate a variety of environmental factors (CPU features)
The content you are editing has changed. Please copy your edits and refresh the page.
Short Description
As an Antelope Core developers, I want to be able to run an automated replay test from genesis within 1 day, to confirm that consensus is maintained and uptime reliability continues.
What
To verify that the consensus mechanism is intact, we can run a full replay from genesis and confirm that every block can be applied without error, and that we achieve the expected state after replay. It may be valuable to run multiple replay tests with different options (for example: supported CPU features, using various WASM runtimes). This would allow us to gain confidence that consensus is robust under many different environmental situations.
Why
Running a full replay from genesis manually would be a months-long endeavor. We need to be able to run a replay test within 1 day in an automated and repeatable fashion.
Considerations
Requirements
Non-requirements (Potential Future Version of Chicken Dance)
Tasks
The text was updated successfully, but these errors were encountered: