Open petermetz opened 3 years ago
Fortunately I didn't face it for 3 runs.
@jagpreetsinghsasan Yeah that's what makes it especially hard to debug, that it works around 9 times out of 10. My best idea so far to make it easier to reproduce is to simulate the constraint hardware resources of the CI environment with a custom dev container for which I have a branch that I've been carrying around (e.g. rebasing onto upstream/main) for just the purpose of trying to reproduce CI-only flaky test issues:
This likely has the same solution as https://github.com/hyperledger/cactus/issues/1469
Describe the bug
We've got a flaky one at:
packages/cactus-plugin-ledger-connector-corda/src/test/typescript/integration/deploy-cordapp-jars-to-nodes-v4.7.test.ts
One of the corda nodes within the container occasionally don't boot up (hangs and then the test waits for an hour before giving up).
To Reproduce
Keep running the test and eventually you'll see it.
Expected behavior
Tests are stable, not flaky.
Logs/Stack traces
The CI logs from the GHA runner as of this morning: logs_10174.zip
Cloud provider or hardware configuration:
GHA runner
Operating system name, version, build:
Ubuntu 20.04
Hyperledger Cactus release version or commit (git rev-parse --short HEAD):
1.0.0-rc.1
Hyperledger Cactus Plugins/Connectors Used
Corda
Additional context
I'll send a separate PR that skips this test to avoid flakiness in the meantime.