✨ e2e: isolate tests with per-scenario dynamic catalogs #2651
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,118 @@ | ||
| # E2E Test Isolation: Per-Scenario Catalogs via Dynamic OCI Image Building | ||
|
|
||
| ## Problem | ||
|
|
||
| E2E test scenarios previously shared cluster-scoped resources (ClusterCatalogs, CRDs, packages), | ||
| causing cascading failures when one scenario left state behind. Parallelism was impossible because | ||
| scenarios conflicted on shared resource names. | ||
|
|
||
| ## Solution | ||
|
|
||
| Each scenario dynamically builds and pushes its own bundle and catalog OCI images at test time, | ||
| parameterized by scenario ID. All cluster-scoped resource names include the scenario ID, making | ||
| conflicts structurally impossible. | ||
|
|
||
| ``` | ||
| Scenario starts | ||
| -> Generate parameterized bundle manifests (CRD names, deployments, etc. include scenario ID) | ||
| -> Build + push bundle OCI images to e2e registry via go-containerregistry | ||
|
Member
Author
There was a problem hiding this comment. I've got this PR updated to keep around the essentials needed to avoid (hopefully) avoid breaking downstream. I've also got a new downstream PR that contains no downstream changes (just cherry-pick from upstream, resolve conflicts, and vendor): openshift/operator-framework-operator-controller#700
Member
Author
|
||
| -> Generate FBC catalog config referencing those bundle image refs | ||
| -> Build + push catalog OCI image to e2e registry | ||
| -> Create ClusterCatalog pointing at the catalog image | ||
| -> Run scenario steps | ||
| -> Cleanup all resources (including catalog) | ||
| ``` | ||
|
|
||
| ### Key Properties | ||
|
|
||
| - Every cluster-scoped resource name includes the scenario ID -- no conflicts by construction. | ||
| - Failed scenario state is preserved for debugging without affecting other scenarios. | ||
| - Parallelism (`Concurrency > 1`) is safe without further changes. | ||
| - Adding new scenarios requires zero coordination with existing ones. | ||
|
|
||
| ## Builder API (`test/e2e/catalog/`) | ||
|
|
||
| Bundles are defined as components of a catalog. A single `Build()` call builds and pushes | ||
| all bundle images, generates the FBC, and pushes the catalog image: | ||
|
|
||
| ```go | ||
| cat := catalog.NewCatalog("test", scenarioID, | ||
| catalog.WithPackage("test", | ||
| catalog.Bundle("1.0.0", catalog.WithCRD(), catalog.WithDeployment(), catalog.WithConfigMap()), | ||
| catalog.Bundle("1.2.0", catalog.WithCRD(), catalog.WithDeployment()), | ||
| catalog.Channel("beta", catalog.Entry("1.0.0"), catalog.Entry("1.2.0")), | ||
| ), | ||
| ) | ||
| result, err := cat.Build(ctx, "v1", localRegistry, clusterRegistry) | ||
| // result.CatalogName = "test-catalog-{scenarioID}" | ||
| // result.CatalogImageRef = "{clusterRegistry}/e2e/test-catalog-{scenarioID}:v1" | ||
| // result.PackageNames = {"test": "test-{scenarioID}"} | ||
| ``` | ||
|
|
||
| ### Bundle Options | ||
|
|
||
| - `WithCRD()` -- CRD with group `e2e-{id}.e2e.operatorframework.io` | ||
| - `WithDeployment()` -- Deployment named `test-operator-{id}` (includes CSV, script ConfigMap, NetworkPolicy) | ||
| - `WithConfigMap()` -- additional test ConfigMap | ||
| - `WithInstallMode(modes...)` -- sets supported install modes on the CSV | ||
| - `WithLargeCRD(fieldCount)` -- CRD with many fields for large bundle testing | ||
| - `WithClusterRegistry(host)` -- overrides the cluster-side registry host (for mirror testing) | ||
| - `StaticBundleDir(dir)` -- reads pre-built bundle manifests without parameterization (e.g. webhook-operator) | ||
| - `BadImage()` -- uses an invalid container image to trigger ImagePullBackOff | ||
| - `WithBundleProperty(type, value)` -- adds a property to bundle metadata | ||
|
|
||
| ## Feature File Conventions | ||
|
|
||
| Feature files define catalogs inline via data tables: | ||
|
|
||
| ```gherkin | ||
| Background: | ||
| Given OLM is available | ||
| And an image registry is available | ||
| And a catalog "test" with packages: | ||
| | package | version | channel | replaces | contents | | ||
| | test | 1.0.0 | alpha | | CRD, Deployment, ConfigMap | | ||
| | test | 1.0.1 | alpha | 1.0.0 | CRD, Deployment, ConfigMap | | ||
| | test | 1.2.0 | beta | | CRD, Deployment | | ||
| ``` | ||
|
|
||
| ### Variable Substitution | ||
|
|
||
| Templates in feature file YAML use these variables: | ||
|
|
||
| | Variable | Expansion | Example | | ||
| |----------|-----------|---------| | ||
| | `${NAME}` | ClusterExtension name | `ce-abc123` | | ||
| | `${TEST_NAMESPACE}` | Scenario namespace | `ns-abc123` | | ||
| | `${SCENARIO_ID}` | Unique scenario identifier | `abc123` | | ||
| | `${PACKAGE:<name>}` | Parameterized package name | `test-abc123` | | ||
| | `${CATALOG:<name>}` | ClusterCatalog resource name | `test-catalog-abc123` | | ||
| | `${COS_NAME}` | ClusterObjectSet name | `cos-abc123` | | ||
|
|
||
| ### Naming Conventions | ||
|
|
||
| | Resource | Pattern | | ||
| |----------|---------| | ||
| | CRD group | `e2e-{id}.e2e.operatorframework.io` | | ||
| | Deployment | `test-operator-{id}` | | ||
| | Package name (FBC) | `{package}-{id}` | | ||
| | Bundle image | `{registry}/bundles/{package}-{id}:v{version}` | | ||
| | Catalog image | `{registry}/e2e/{name}-catalog-{id}:{tag}` | | ||
| | ClusterCatalog | `{name}-catalog-{id}` | | ||
| | Namespace | `ns-{id}` | | ||
| | ClusterExtension | `ce-{id}` | | ||
|
|
||
| ## Registry Access | ||
|
|
||
| An in-cluster OCI registry (`test/internal/registry/`) stores bundle and catalog images. | ||
| The registry runs as a ClusterIP Service; there is no NodePort or kind `extraPortMappings`. | ||
|
|
||
| The test runner reaches the registry via **Kubernetes port-forward** (SPDY through the API | ||
| server), which works regardless of the cluster's network topology. A `sync.OnceValues` in the | ||
| step definitions starts the port-forward once and returns the dynamically assigned | ||
| `localhost:<port>` address used for all `crane.Push` / `crane.Tag` calls. | ||
|
|
||
| In-cluster components (e.g. the catalog unpacker) pull images using the Service DNS name | ||
| (`docker-registry.operator-controller-e2e.svc.cluster.local:5000`), resolved by CoreDNS. | ||
| Containerd on the node is never involved because the registry only holds OCI artifacts | ||
| consumed by Go code, not container images for pods. | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Uh oh!
There was an error while loading. Please reload this page.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@joelanford should we get it merged?
It might be nice have the design of the tests but should this dir be the correct one/
I think it was generated by CLAUDE to address the need.
Uh oh!
There was an error while loading. Please reload this page.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do we have a directory for this kind of thing? I do think we should merge this kind of doc, and I was definitely intending to have this exact discussion.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If we want define how our e2e test is design I would put that in the Contributing.MD
If you want to track design decisions then we need a new dir, so far all is in the google drive.