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
This happens rarely, and on re-running the pipeline is usually passes.
Why is this issue relevant?
We want CI to pass consistently
What steps are required to resolve this?
My hunch was that it was related to a timeout for the http request made by the propagation pallet.
From a message in discord:
regarding the test_store_share test which occasionally fails that i was going on about. I think it might be related to the http request 'deadline' of 2 seconds set in the propagation pallet. The user/new endpoint doesn't give a response until after DKG is completed, and it might be that if the request gets aborted because of this timeout, axum stops the running task. Im not totally sure i've got this right, but since it also effects the test im currently writing it will hopefully become clearer whats going on and i'll put up a card explaining what needs fixing.
let deadline = sp_io::offchain::timestamp().add(Duration::from_millis(2_000));
However I did at some point try increasing this an still had this problem, so it could be something else. I will try to find the discussion on that, i think it was in a PR.
Does this change the spec? HTTP, extrinsic, or storage? Is it breaking? Clearly describe the new interface.
I think it should be possible to fix this without a breaking change.
The text was updated successfully, but these errors were encountered:
This is from a conversation on discord in September that i should probably have made into an issue then.
Occasionally the test_store_share test fails in CI. Here is an example: https://app.circleci.com/pipelines/github/entropyxyz/entropy-core/1627/workflows/1d989818-e294-4a2e-990c-09a774abdadd/jobs/6377
I will try to find a more recent example.
This happens rarely, and on re-running the pipeline is usually passes.
Why is this issue relevant?
We want CI to pass consistently
What steps are required to resolve this?
My hunch was that it was related to a timeout for the http request made by the propagation pallet.
From a message in discord:
Here is where that timeout gets set:
entropy-core/pallets/propagation/src/lib.rs
Line 71 in 357477b
However I did at some point try increasing this an still had this problem, so it could be something else. I will try to find the discussion on that, i think it was in a PR.
Does this change the spec? HTTP, extrinsic, or storage? Is it breaking? Clearly describe the new interface.
I think it should be possible to fix this without a breaking change.
The text was updated successfully, but these errors were encountered: