We do not systematically RollForward
First step to roll backward in ModelSpec is to split blocks generation from rolling forward.
First step to roll backward in ModelSpec is to split blocks generation from rolling forward.
This will store and all hydra-cluster state directories at /tmp/hydra-cluster-e2e-*, which stay for failing tests. The Github workflow then can pick these up to provide hydra-node and cardano-node logs for failing e2e tests in CI.
To show the chain slot and block header hash on which the Head got Init.
We change this to make it more robust to open heads. Now, all head participants would try to open a head with a collect, not only the last to commit. This is also how it is written in the spec right now.
This does not use the testnet magic and works slightly differently.
Next step will be responsible for merging released and unstable docs.
This will help later when versioning docs, because different versions will have different base url.
This will store and all hydra-cluster state directories at /tmp/hydra-cluster-e2e/, which stay for failing tests. The Github workflow then can pick these up to provide hydra-node and cardano-node logs for failing e2e tests in CI.
Play with runners
main-net has mem units: 14000000
We remove the cachix step because: 1. we encounter a bunch of troubles with making it work on a self-hosted runner 2. we don't care about uploading built stuff to cachix from the smoke tests, there are other C.I. steps to do it 3. downloading from cachix works anyway We remove the github cache because we're going to run the smoke tests several times on the same self-hosted runner so no need for a cache as everything is already available locally. We use a shared absolute path to store the cardano database. We should maybe move the hydra state to a more local and voltaile directory since we remove it everytime but that's good enough for now. We ensure that we do not share the same cardano databases storage between two concurrent work with concurrency github parameter.
Next step will be responsible for merging released and unstable docs.
This will produce a single docs build combining both released and unstable. This workflow will only get triggered if CI completed with success. Currently it is only uploading the artifact result. Soon it will publish docs to gh-pages.
This will help later when versioning docs, because different versions will have different base url.