Skip to content

Conversation

@lrubasze
Copy link
Contributor

@lrubasze lrubasze commented Jun 4, 2025

This PR re-enables Zombienet CI tests.

Changes:

  • Enable Zombienet Polkadot and Substrate tests
  • Zombienet Cumulus will be enabled separately, when being migrated to zombienet-sdk
  • Switch zombienet from k8s to native provider
    k8s turned out to be unstable for both zombienet and zombienet-sdk. Issues observed:
    • problem with launching a pod
    • pods (and thus nodes) were not spawned at the same time (differences up to 120s), which affected some tests
    • kubectl command failed
      Observed many times for:
      • kubectl cp
      • kubectl exec
      • kubectl logs
        If we ever want to switch back to k8s we must ensure above issues no longer exist.
  • Tweaks some tests to make sure they constantly pass
    Authors of those tests are kindly asked to review the changes.
  • Some improvements and fixes zombienet and zombienet-sdk frameworks
  • Assign more beefy runners for more demanding tests
    Rule of thumb: use large runner if spawned network consist of more than 4 nodes
  • Disable some tests to let their authors to stabilize them
    Created a dedicated .github/zombienet-flaky-tests file for more clarity
    ATM there are 16 flaky tests. Their authors are kindly asked to fix them and re-enable.

@paritytech-review-bot paritytech-review-bot bot requested a review from a team June 24, 2025 12:54
Copy link
Contributor

@alvicsam alvicsam left a comment

Choose a reason for hiding this comment

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

Amazing job!
I have only a few comments related to inputs of a reusable github action.

@paritytech-review-bot paritytech-review-bot bot requested a review from a team June 24, 2025 16:44
@lrubasze
Copy link
Contributor Author

Test results from nightly test 2025-06-24

  • Polkadot
    Failed / total runs: 0 / 54 (failure rate 0%)
    Skipped (flaky) / Total test cnt: 10 / 32
  • Substrate
    Failed / total runs: 0 / 101 (failure rate 0%)
    Skipped (flaky) / Total test cnt: 2 / 4

image

@paritytech-review-bot paritytech-review-bot bot requested a review from a team June 25, 2025 07:56
Copy link
Contributor

@alindima alindima left a comment

Choose a reason for hiding this comment

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

Only some nits remaining from me

# We need to use an older version of polkadot/polkadot-parachain, particulary a version that doesn't includes
# https://github.com/paritytech/polkadot-sdk/pull/1644, and since this change is also used by the collators we need to
# keep this test to ensure that works with older versions.
- name: download_old_release_bins
Copy link
Contributor

Choose a reason for hiding this comment

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

could we maybe abstract this into some script? I'm pretty sure other tests will need it too (and it's already duplicated for 2 tests)

Copy link
Contributor

Choose a reason for hiding this comment

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

We could do this later if you want, I really want us to start running zombienet tests again

Copy link
Contributor

Choose a reason for hiding this comment

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

Yes, is alredy used twice. I will move the logic to a reusable action. Thx!!

@pepoviola pepoviola added this pull request to the merge queue Jun 26, 2025
Merged via the queue into master with commit dbfed5a Jun 26, 2025
282 of 283 checks passed
@pepoviola pepoviola deleted the ci/make-zombienet-ci-great-again branch June 26, 2025 10:59
alvicsam pushed a commit that referenced this pull request Oct 17, 2025
Closes #8871
Related to #8748

Fix `zombienet-substrate-0002-validators-warp-sync`
alvicsam pushed a commit that referenced this pull request Oct 17, 2025
This PR re-enables Zombienet CI tests.


Changes:
- Enable Zombienet Polkadot and Substrate tests
- Zombienet Cumulus will be enabled separately, when being migrated to
`zombienet-sdk`
- Switch zombienet from `k8s` to `native` provider
`k8s` turned out to be unstable for both `zombienet` and
`zombienet-sdk`. Issues observed:
  - problem with launching a pod
- pods (and thus nodes) were not spawned at the same time (differences
up to 120s), which affected some tests
  - `kubectl` command failed 
  Observed many times for:
    - kubectl cp
    - kubectl exec
    - kubectl logs
If we ever want to switch back to `k8s` we must ensure above issues no
longer exist.
- Tweaks some tests to make sure they constantly pass
Authors of those tests are kindly asked to review the changes.
- Some improvements and fixes `zombienet` and `zombienet-sdk` frameworks
- Assign more beefy runners for more demanding tests
Rule of thumb: use large runner if spawned network consist of more than
4 nodes
- Disable some tests to let their authors to stabilize them
Created a dedicated `.github/zombienet-flaky-tests` file for more
clarity
ATM there are 16 flaky tests. Their authors are kindly asked to fix them
and re-enable.

---------

Co-authored-by: Javier Viola <[email protected]>
Co-authored-by: Javier Viola <[email protected]>
Co-authored-by: cmd[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Anthony Lazam <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

R0-no-crate-publish-required The change does not require any crates to be re-published. T10-tests This PR/Issue is related to tests. T18-zombienet_tests Trigger zombienet CI tests.

Projects

None yet

Development

Successfully merging this pull request may close these issues.