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
Our smoke-tests run on a custom github runner we have setup in aws and while running the smoke-tests is easy - making sure they work on all networks is a bit of a problem. In my experience we always loose at least a bit of time before each release on this specific problem.
We can choose to use mithril to download the cardano-node database. Once the db is downloaded, on mainnet it takes a long time to re-validate all blocks and start the node. This usually leads to timeout in the github actions. If the process fails after publishing the Hydra scripts and before successful termination we also loose money! (this is because we put the link to the smoke-test we ran in release notes)
How to improve upon this situation?
One option would be to use blockfrost and we already have a dedicated item for integration Blockfrost chain layer #1305
Investigate possible improvements for making sure cardano-node runs as fast as possible after mithril download (skip validating blocks)
Don't run smoke-tests at all? We can do manual runs locally which would save us time it takes to figure out what is happening on remote machine.
Why
Our smoke-tests run on a custom github runner we have setup in aws and while running the smoke-tests is easy - making sure they work on all networks is a bit of a problem. In my experience we always loose at least a bit of time before each release on this specific problem.
We can choose to use mithril to download the cardano-node database. Once the db is downloaded, on mainnet it takes a long time to re-validate all blocks and start the node. This usually leads to timeout in the github actions. If the process fails after publishing the Hydra scripts and before successful termination we also loose money! (this is because we put the link to the smoke-test we ran in release notes)
How to improve upon this situation?
One option would be to use blockfrost and we already have a dedicated item for integration Blockfrost chain layer #1305
Investigate possible improvements for making sure cardano-node runs as fast as possible after mithril download (skip validating blocks)
Don't run smoke-tests at all? We can do manual runs locally which would save us time it takes to figure out what is happening on remote machine.
Example: https://github.com/cardano-scaling/hydra/actions/runs/13115950184/job/36591124801
This test was red because of
ConnectionClosed
error while locally it worked just fine.The text was updated successfully, but these errors were encountered: