Skip to content

packaging: harden smoke-test cluster startup#458

Open
dkropachev wants to merge 1 commit into
scylladb:masterfrom
dkropachev:packaging/harden-smoke-test-cluster-startup
Open

packaging: harden smoke-test cluster startup#458
dkropachev wants to merge 1 commit into
scylladb:masterfrom
dkropachev:packaging/harden-smoke-test-cluster-startup

Conversation

@dkropachev
Copy link
Copy Markdown
Contributor

@dkropachev dkropachev commented May 22, 2026

Summary

  • harden the non-Windows smoke-test cluster startup path used by package tests
  • retry docker compose up, wait for the Scylla container to become healthy, and dump compose state/logs when startup fails
  • keep compose cleanup non-fatal so transient teardown errors do not hide the real startup failure

Testing

  • not run locally on macOS in this branch split; logic was originally validated on the combined branch via make test-package-linux and by the succeeding macOS package CI job after the hardening was introduced

Checklist

  • I have split my patch into logically separate commits.
  • All commit messages clearly explain what they change and why.
  • PR description sums up the changes and reasons why they should be introduced.
  • I have implemented Rust unit tests for the features/changes introduced.
  • I have enabled appropriate tests in Makefile in {SCYLLA,CASSANDRA}_(NO_VALGRIND_)TEST_FILTER.
  • I added appropriate Fixes: annotations to PR description.

@dkropachev dkropachev marked this pull request as ready for review May 22, 2026 16:41
Harden the non-Windows smoke-test path that starts the Scylla package-test cluster. Keep the existing docker-compose-based flow, but stop assuming that `docker compose up --wait` is enough to guarantee that the cluster is actually ready for the smoke binary.

Reuse a dedicated compose-file variable, make teardown non-fatal, retry startup a few times, and inspect the Scylla container health before running the smoke app. When startup still fails, dump compose state and recent Scylla logs so CI failures are diagnosable from the job output.
@dkropachev dkropachev force-pushed the packaging/harden-smoke-test-cluster-startup branch from f45e748 to e23f18f Compare May 22, 2026 16:43
@dkropachev dkropachev requested a review from wprzytula May 22, 2026 16:47
Copy link
Copy Markdown
Contributor

@wprzytula wprzytula left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I doubt this is needed at all. See docker/compose#8777 (comment). If a healthcheck condition is defined (which is the case in our docker-compose.yml), docker compose up --wait waits until the service is healthy.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants