Enable disableMDevConfiguration feature gate before vGPU setup#4959
Merged
Conversation
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: ASSERTIVE Plan: Pro Run ID: 📒 Files selected for processing (2)
📝 WalkthroughWalkthroughThe PR centralizes ChangesDISABLE_MDEV_CONFIGURATION + vGPU setup refactor
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Suggested reviewers
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Warning Review ran into problems🔥 ProblemsLinked repositories: Your configuration references 1 linked repositories, but your current plan allows 0. Analyzed ``, skipped Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Contributor
|
Report bugs in Issues Welcome! 🎉This pull request will be automatically processed with the following features: 🔄 Automatic Actions
📋 Available CommandsPR Status Management
Review & Approval
Testing & Validation
Container Operations
Cherry-pick Operations
Label Management
✅ Merge RequirementsThis PR will be automatically approved when the following conditions are met:
📊 Review ProcessApprovers and ReviewersApprovers:
Reviewers:
Available Labels
AI Features
💡 Tips
For more information, please refer to the project documentation or contact the maintainers. |
openshift-virtualization-qe-bot
added
the
tox:verify-tc-requirement-polarion:running
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #4959 +/- ##
=======================================
Coverage 98.67% 98.67%
=======================================
Files 25 42 +17
Lines 2487 2493 +6
=======================================
+ Hits 2454 2460 +6
Misses 33 33
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
openshift-virtualization-qe-bot-2
requested review from
RoniKishner,
hmeir,
rlobillo and
rnetser
openshift-virtualization-qe-bot
added
tox:verify-tc-requirement-polarion:passed
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@tests/virt/conftest.py`:
- Around line 243-264: The two fixtures hco_with_disable_mdev_configuration and
gpu_nodes_labeled_with_vgpu_config lack type hints; add explicit typing for
their parameters and return types (fixtures that yield should be typed as
typing.Generator[None, None, None] or typing.Iterator[None]). Annotate
hco_with_disable_mdev_configuration's parameter
hyperconverged_resource_scope_session and set the fixture return type to
Generator[None, None, None]; for gpu_nodes_labeled_with_vgpu_config, annotate
hco_with_disable_mdev_configuration, nodes_with_supported_gpus, and
supported_gpu_device with their concrete fixture types (or typing.Any if the
concrete types are defined elsewhere) and give the function a Generator[None,
None, None] return type. Ensure to import typing.Generator (or typing.Iterator)
and any concrete fixture types used.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: ASSERTIVE
Plan: Pro
Run ID: 06180790-df87-440b-8590-cda39b6df430
📒 Files selected for processing (3)
tests/install_upgrade_operators/constants.pytests/virt/conftest.pyutilities/constants.py
📜 Review details
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: can-be-merged
🧰 Additional context used
📓 Path-based instructions (4)
**/*.py
📄 CodeRabbit inference engine (AGENTS.md)
**/*.py: Never add linter suppressions like# noqa,# type: ignore, or# pylint: disable. Fix the code instead. If you believe a rule is wrong, ask the user for explicit approval.
Search the codebase for existing implementations before writing new code. Checkutilities/,libs/,tests/, andpyproject.tomldependencies. Never duplicate logic—extract to shared modules. Usepyhelper_utils.shell.run_commandfor shell commands instead ofsubprocess.run, and useocp-resourcesclasses instead of raw YAML dicts.
Type hints are MANDATORY. Use mypy strict mode inlibs/and all new public functions under utilities. UseTYPE_CHECKINGfor type-only imports to avoid runtime overhead and circular imports.
Write Google-format docstrings for all public functions with non-obvious return values or side effects.
Always useuv runto execute commands. Never executepython,pip,pytest,tox, orpre-commitdirectly. Useuv run python,uv run pytest,uv run tox,uv run pre-commit, anduv addfor package installation.
Always use absolute imports. Never use relative imports.
Prefer specific imports usingfrom module import funcfor functions and constants. Usefrom package import module(thenmodule.Name) when retaining the module name meaningfully improves readability. Never use bareimport modulewithout afromclause.
Always use named arguments for function calls with more than one argument.
Never use single-letter variable names. Always use descriptive, meaningful names.
No dead code. Every function, variable, and fixture must be used or removed. Code marked with# skip-unused-codeis excluded from dead code analysis (enforced via custom ruff plugin).
Prefer direct attribute access usingfoo.attr. Save to variables only when reusing the same attribute multiple times improves readability or extracting clarifies intent.
Imports must always be at the top of the module. Do not import inside functions.
No defensive programming. Fail...
Files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
**
⚙️ CodeRabbit configuration file
**: ## PR Template Validation
Check the PR description for required sections from.github/pull_request_template.md.
Required sections (must be present, even if empty):
##### What this PR does / why we need it:— MUST be present AND have meaningful content.
Flag as HIGH if the section is missing, empty, whitespace-only, contains only HTML comments,
or contains only placeholder tokens such asTBD,TBA,N/A,-,—,none, or..##### Which issue(s) this PR fixes:— must be present (may be empty)##### Special notes for reviewer:— must be present (may be empty)##### jira-ticket:— must be present (may be empty)
If any required section is absent, orWhat this PR does / why we need it:has no content,
flag it as HIGH severity and ask the author to restore the missing template section(s).
Files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
tests/**/*.py
📄 CodeRabbit inference engine (AGENTS.md)
Use pytest assertions:
assert actual == expected. Never useself.assertEqual(). Include failure messages:assert condition, 'descriptive message explaining failure'.
Files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
**/conftest.py
📄 CodeRabbit inference engine (AGENTS.md)
**/conftest.py:conftest.pyis for fixtures only. Helper functions, utility functions, and classes must NOT be defined in conftest.py or test_*.py. Place them in dedicated utility modules instead.
Fixtures MUST do ONE action only (single responsibility). Use NOUNS for naming (what they provide), NEVER verbs. Correct:vm_with_disk. Incorrect:create_vm_with_disk. Userequest.paramwith dict structure for complex parameters. Order pytest native fixtures first, then session-scoped, then module/class/function scoped. Fixtures should return/yield the resource even if it's a setup fixture. Scope rules:scope='function'(default) for test isolation,scope='class'for setup shared across test class,scope='module'for expensive setup in a test module,scope='session'for setup persisting entire test run. Never use broader scope if fixture modifies state or creates per-test resources.
Files:
tests/virt/conftest.py
🧠 Learnings (67)
📚 Learning: 2026-01-12T11:24:13.825Z
Learnt from: servolkov
Repo: RedHatQE/openshift-virtualization-tests PR: 3387
File: tests/network/provider_migration/libprovider.py:50-52
Timestamp: 2026-01-12T11:24:13.825Z
Learning: In the RedHatQE/openshift-virtualization-tests repository, when catching exceptions in Python, use LOGGER.error before re-raising and do not replace it with LOGGER.exception in except blocks. This follows the established pattern across the codebase.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-12T14:25:05.723Z
Learnt from: rnetser
Repo: RedHatQE/openshift-virtualization-tests PR: 3366
File: tests/storage/cdi_clone/test_clone.py:5-9
Timestamp: 2026-01-12T14:25:05.723Z
Learning: In Python tests and utility code across the repository, bitmath.parse_string_unsafe correctly parses Kubernetes quantities (e.g., '4Gi', '512Mi', PVC storage requests) without supplying system=bitmath.NIST. There are 30+ usages indicating this is the standard behavior. Reviewers should verify that code that builds or compares quantity strings does not pass the NIST parameter, and if a new test relies on quantity parsing, assume no NIST parameter is required unless explicitly documented.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-20T01:03:13.139Z
Learnt from: servolkov
Repo: RedHatQE/openshift-virtualization-tests PR: 3387
File: tests/network/provider_migration/libprovider.py:1-8
Timestamp: 2026-01-20T01:03:13.139Z
Learning: In the openshift-virtualization-tests repository, Python imports should consistently use module-level imports for the logging module (i.e., import logging) rather than from logging import ... The established pattern spans 270+ files and should not be flagged for refactoring. Apply this guideline to Python files across the repo (e.g., tests/network/provider_migration/libprovider.py).
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-21T21:26:41.805Z
Learnt from: geetikakay
Repo: RedHatQE/openshift-virtualization-tests PR: 3559
File: utilities/infra.py:251-254
Timestamp: 2026-01-21T21:26:41.805Z
Learning: In the RedHatQE/openshift-virtualization-tests repository, when reviewing Python code, recognize that with Python 3.14 the syntax 'except ValueError, TypeError:' is valid if there is no 'as' clause, and should not be flagged as Python 2 syntax. If you use an 'as' binding (e.g., 'except (ValueError, TypeError) as e:'), parentheses are required. Ensure this pattern is version-consistent and not flagged as Python 2 syntax when 'as' is absent.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-25T13:18:21.675Z
Learnt from: jpeimer
Repo: RedHatQE/openshift-virtualization-tests PR: 3571
File: tests/storage/storage_migration/utils.py:158-167
Timestamp: 2026-01-25T13:18:21.675Z
Learning: In reviews of the openshift-virtualization-tests repo (and similar Python code), avoid suggesting minor stylistic changes that require extra verification (e.g., removing dict.keys() checks for membership) unless the change has clear correctness or maintainability impact. Focus on fixes with observable behavior, security, performance, or maintainability benefits; defer low-impact style tweaks that are costly to verify.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-02-18T06:35:39.536Z
Learnt from: rnetser
Repo: RedHatQE/openshift-virtualization-tests PR: 3847
File: utilities/virt.py:2449-2453
Timestamp: 2026-02-18T06:35:39.536Z
Learning: In Python code, a function named clearly and self-descriptively can be deemed not to require a docstring. However, treat this as a context-specific guideline and not a universal rule. For public APIs or functions with side effects, prefer concise docstrings explaining behavior, inputs, outputs, and side effects. This guidance is based on the example in utilities/virt.py from RedHatQE/openshift-virtualization-tests where validate_libvirt_persistent_domain(vm, admin_client) was considered self-documenting.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-02-23T16:33:22.070Z
Learnt from: vsibirsk
Repo: RedHatQE/openshift-virtualization-tests PR: 3883
File: utilities/pytest_utils.py:441-463
Timestamp: 2026-02-23T16:33:22.070Z
Learning: In Python code reviews, the guideline to always use named arguments for multi-argument calls does not apply to built-ins or methods that have positional-only parameters (those defined with a / in their signature). Do not flag or require named arguments for calls like dict.get(key, default=None, /), list.pop(), str.split(sep, maxsplit) and similar built-ins that cannot accept keyword arguments. Apply the named-argument rule only to functions/methods that explicitly accept keyword arguments.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-03-17T01:32:02.617Z
Learnt from: dshchedr
Repo: RedHatQE/openshift-virtualization-tests PR: 4118
File: utilities/database.py:0-0
Timestamp: 2026-03-17T01:32:02.617Z
Learning: In RedHatQE/openshift-virtualization-tests, when reviewing Python files, post targeted inline comments on the Files changed tab at the exact location (file and line) of the issue rather than opening a single discussion thread for multiple issues. This should be done for each applicable location to improve traceability and clarity. If multiple issues exist in the same file, address them with separate inline comments pointing to the specific lines.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-03-17T01:32:02.617Z
Learnt from: dshchedr
Repo: RedHatQE/openshift-virtualization-tests PR: 4118
File: utilities/database.py:0-0
Timestamp: 2026-03-17T01:32:02.617Z
Learning: In the RedHatQE/openshift-virtualization-tests repository, CodeRabbit should post targeted inline comments at each applicable location in the Files Changed tab, rather than aggregating multiple issues into a single PR discussion thread reply. This guideline applies to all Python files (any file ending in .py) changed in a PR; for non-Python files, follow the same inline-comment-at-location principle if relevant.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-04T13:45:29.122Z
Learnt from: rnetser
Repo: RedHatQE/openshift-virtualization-tests PR: 4725
File: utilities/console.py:54-59
Timestamp: 2026-05-04T13:45:29.122Z
Learning: During review of RedHatQE/openshift-virtualization-tests “lint-cleanup” PRs (e.g., changes targeting lint issues like stale noqa/utf-8 headers), do not flag existing `# type: ignore` directives that were already present before the PR and were not introduced or modified by the PR. Only raise findings for `# type: ignore` suppressions that the PR itself adds, changes, or otherwise makes newly effective (i.e., they appear in the diff as additions/edits).
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-04T13:45:33.892Z
Learnt from: rnetser
Repo: RedHatQE/openshift-virtualization-tests PR: 4725
File: tests/virt/cluster/common_templates/centos/test_centos_os_support.py:78-83
Timestamp: 2026-05-04T13:45:33.892Z
Learning: When reviewing lint-cleanup or formatting-only pull requests in this repo (e.g., changes like removing/updating `# noqa` comments or UTF-8 headers), do not raise findings for code patterns that already existed before the PR. Specifically, if a problematic construct such as `.is_connective(tcp_timeout=120)` was present in the base branch, suppress that finding and only raise issues when the PR itself introduces or modifies that construct (i.e., the diff adds/changes the call or its arguments). Apply this rule across all Python files (`**/*.py`).
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-05T17:01:15.294Z
Learnt from: dshchedr
Repo: RedHatQE/openshift-virtualization-tests PR: 4739
File: tests/virt/node/descheduler/conftest.py:2-2
Timestamp: 2026-05-05T17:01:15.294Z
Learning: In this repo’s Python code, it’s acceptable (and preferred by convention) to build `run_command` inputs using `shlex.split(f"<command> {arg}")` rather than converting to direct list literals like `['oc', 'adm', 'uncordon', name]`. During code review, generally don’t flag `shlex.split(...)` usage for `run_command` calls and don’t suggest replacing it with list literals; the string-form pattern is used to keep commands readable and consistent with how they’re typed in a terminal.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-08T12:49:20.694Z
Learnt from: geetikakay
Repo: RedHatQE/openshift-virtualization-tests PR: 4788
File: utilities/os_utils.py:257-262
Timestamp: 2026-05-08T12:49:20.694Z
Learning: In RedHatQE/openshift-virtualization-tests, the Ruff flake8-boolean-trap rules FBT001/FBT002 are intentionally not enabled (pyproject.toml does not select the FBT rules; confirmed via `ruff check --show-settings`). Therefore, do not flag boolean positional parameters as FBT001/FBT002 violations in this repository. If Ruff configuration changes and starts selecting FBT rules, this exception should be reconsidered.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-12T05:10:24.601Z
Learnt from: acinko-rh
Repo: RedHatQE/openshift-virtualization-tests PR: 4780
File: tests/storage/utils.py:568-572
Timestamp: 2026-05-12T05:10:24.601Z
Learning: In this repository, Ruff rule UP043 ("unnecessary default type arguments") is enforced. When annotating `collections.abc.Generator` return types, prefer the single-parameter form `Generator[YieldType]` rather than `Generator[YieldType, None, None]`. Explicit `None, None` for the SendType and ReturnType are unnecessary defaults (per PEP 696) and will trigger UP043. Apply this consistently across all Python files.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-13T19:23:09.603Z
Learnt from: Anatw
Repo: RedHatQE/openshift-virtualization-tests PR: 4833
File: tests/network/localnet/migration_stuntime/libstuntime.py:25-25
Timestamp: 2026-05-13T19:23:09.603Z
Learning: In this repository, do not recommend adding `from __future__ import annotations` to fix forward-reference type annotation issues (e.g., Ruff UP037). Follow the established convention: use quoted string type annotations for forward references when the referenced class/type is defined later in the same file (e.g., `"ContinuousPing"`), and prefer `typing.Self` for self-referential return types.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-18T06:30:56.781Z
Learnt from: EdDev
Repo: RedHatQE/openshift-virtualization-tests PR: 4819
File: utilities/unittests/test_bitwarden.py:207-207
Timestamp: 2026-05-18T06:30:56.781Z
Learning: During Ruff/lint rule-enablement PRs in this repository (e.g., when introducing a new rule like PLC0415), it’s acceptable to keep CI green by adding per-line, targeted suppressions for pre-existing violations: add only `# noqa: <single-ruff-rule-id>` at the end of the specific violating line. In this PR context, reviewers should NOT flag these targeted `# noqa: PLC0415` comments as policy violations, assuming the suppression is for a pre-existing issue and is documented in the PR description as a candidate for follow-up cleanup. Do not allow blanket `# noqa` (without a specific rule) or `per-file-ignores`; those remain disallowed.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-18T06:31:12.015Z
Learnt from: EdDev
Repo: RedHatQE/openshift-virtualization-tests PR: 4819
File: utilities/unittests/test_pytest_utils.py:270-270
Timestamp: 2026-05-18T06:31:12.015Z
Learning: In RedHatQE/openshift-virtualization-tests, if a PR is a Ruff rule-enforcement PR and its “Special notes for reviewer” documents that pre-existing Ruff violations are being temporarily handled via per-line suppressions (e.g., `# noqa: PLC0415`) to keep CI green, reviewers should treat those specific `# noqa: <rule>` comments as an agreed, temporary mechanism. Do not flag them as code-quality issues and do not recommend removing, consolidating, or refactoring those suppressions within the same PR; cleanup/remediation is expected to happen in dedicated follow-up PRs instead.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-18T06:31:15.083Z
Learnt from: EdDev
Repo: RedHatQE/openshift-virtualization-tests PR: 4819
File: utilities/unittests/test_data_collector.py:304-304
Timestamp: 2026-05-18T06:31:15.083Z
Learning: When reviewing Python code in this repository for Ruff/linter rule rollouts, do not treat temporary suppression comments as violations in the specific migration scenario where a PR enables a new Ruff rule (e.g., PLC0415) and the PR description explicitly documents that all *pre-existing* violations are being annotated with `# noqa: <RULE>` as a short-lived measure. In that case, only flag `# noqa: <RULE>` suppressions that are newly introduced on code that did not previously violate the rule—i.e., verify via the PR diff against the prior state (and/or prior Ruff findings) that the suppressed line was already violating before the rule was enabled. Ignore suppressions that are covering violations that existed before the new rule rollout and were intentionally bulk-added for cleanup in follow-up PRs.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-18T06:31:20.848Z
Learnt from: EdDev
Repo: RedHatQE/openshift-virtualization-tests PR: 4819
File: utilities/unittests/test_hco.py:501-501
Timestamp: 2026-05-18T06:31:20.848Z
Learning: When reviewing Python code in RedHatQE/openshift-virtualization-tests, avoid flagging Ruff `# noqa: <RULE>` suppressions as issues if they were intentionally added as a temporary measure to keep CI green after a PR enables a new Ruff/lint rule (e.g., PLC0415) and the PR description documents this under "Special notes for reviewer". Treat these suppressions as deferred technical debt. Only flag `# noqa: PLC0415` (and similar rule-specific suppressions) when they are newly introduced without an accompanying documented intent in the PR (and thus appear to be masking a new violation rather than a pre-existing one).
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-18T09:09:09.479Z
Learnt from: EdDev
Repo: RedHatQE/openshift-virtualization-tests PR: 4878
File: utilities/unittests/test_pytest_utils.py:2194-2197
Timestamp: 2026-05-18T09:09:09.479Z
Learning: In this repository (RedHatQE/openshift-virtualization-tests), do not flag missing return type annotations or missing argument type annotations as Ruff “ANN” rule violations (e.g., ANN001/ANN002/ANN201/ANN202). The repo’s Ruff configuration does not enable ANN rules and only uses `extend-select = ["PLC0415"]`, so missing type annotations should not be treated as ANN lint failures during code review.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-19T07:48:17.119Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 4784
File: libs/vm/affinity.py:104-104
Timestamp: 2026-05-19T07:48:17.119Z
Learning: When using Kubernetes API models like `NodeSelectorRequirement` or `LabelSelectorRequirement` with operators `Exists` or `DoesNotExist`, the `values` field must not be non-empty. It is valid for `values` to be omitted / left as `None` (Python) / passed as `null`—Kubernetes rejects non-empty `values` for these operators, but does not require the field to be present or explicitly set to an empty list. In code reviews, do not treat missing `values=[]` for `Exists`/`DoesNotExist` as a validation issue; only flag cases where `values` is provided with actual elements.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-19T07:48:17.119Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 4784
File: libs/vm/affinity.py:104-104
Timestamp: 2026-05-19T07:48:17.119Z
Learning: When constructing Kubernetes `NodeSelectorRequirement` (or `LabelSelectorRequirement`) objects in code, do not treat `values` being omitted, `None`, or an empty list as an API-validation problem when the requirement’s operator is `Exists` or `DoesNotExist`. Per the Kubernetes API spec, these operators only require that the `values` array is not non-empty (i.e., it must be empty); they do not require the field to be explicitly present as `[]`. Therefore, reviewers should not flag `values=None`/missing `values` for `Exists`/`DoesNotExist`.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-02-18T06:34:38.042Z
Learnt from: rnetser
Repo: RedHatQE/openshift-virtualization-tests PR: 3847
File: tests/virt/cluster/common_templates/utils.py:58-58
Timestamp: 2026-02-18T06:34:38.042Z
Learning: In RedHatQE/openshift-virtualization-tests, treat 'public' functions as those defined in any Python files under libs/ or utilities/ (any depth). Functions inside nested test directories (e.g., tests/virt/cluster/common_templates/, tests/virt/node/, etc.) are test helpers and do not require Google-format docstrings unless explicitly requested. Use this rule during reviews to decide whether to enforce docstrings on public API functions in libs/utilities.
Applied to files:
utilities/constants.py
📚 Learning: 2026-05-03T15:38:09.624Z
Learnt from: rnetser
Repo: RedHatQE/openshift-virtualization-tests PR: 4701
File: tests/virt/node/general/test_windows_vtpm_bitlocker.py:50-52
Timestamp: 2026-05-03T15:38:09.624Z
Learning: During review of PRs that are lint cleanups or tooling/version bumps, do not flag code-quality issues for patterns that pre-existed before the PR. Specifically, if the diff does not introduce/modify constructs such as nested `if` blocks or unnecessary list comprehensions, treat them as known/deferred and leave them for dedicated follow-up cleanup PRs. Only raise issues when the PR itself adds, changes, or refactors the problematic code.
Applied to files:
utilities/constants.pytests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-15T18:42:02.504Z
Learnt from: geetikakay
Repo: RedHatQE/openshift-virtualization-tests PR: 4860
File: utilities/hco.py:385-389
Timestamp: 2026-05-15T18:42:02.504Z
Learning: In this repository, under the Python `utilities/` directory (utility/helper modules, not tests), do not flag bare `assert` statements as correctness or style issues. The codebase conventionally uses `assert` in these utilities (e.g., `utilities/virt.py`, `utilities/infra.py`, etc.) and does not enforce running Python with `-O`/`PYTHONOPTIMIZE`, so the usual “asserts may be stripped” concern should not be treated as a review blocker here.
Applied to files:
utilities/constants.py
📚 Learning: 2025-12-15T12:33:06.686Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 3024
File: tests/network/connectivity/utils.py:17-17
Timestamp: 2025-12-15T12:33:06.686Z
Learning: In the test suite, ensure the ipv6_network_data fixture returns a factory function (Callable) and that all call sites invoke it to obtain the actual data dict, i.e., use ipv6_network_data() at call sites. This enables future extensibility for configuring secondary interfaces' IP addresses without changing call sites. Apply this pattern to all Python test files under tests.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-14T04:08:23.032Z
Learnt from: dshchedr
Repo: RedHatQE/openshift-virtualization-tests PR: 3404
File: tests/virt/upgrade/conftest.py:291-301
Timestamp: 2026-01-14T04:08:23.032Z
Learning: In the openshift-virtualization-tests repository, when using VirtualMachine objects from ocp-resources in tests, if vm.ready is True, vm.vmi is guaranteed to exist. Therefore, you can access vm.vmi.instance.status or vm.vmi attributes without additional defensive checks (e.g., if vm.vmi: ...). Do not rely on vm.vmi being present when vm.ready may be False; guard those code paths accordingly. This guideline applies to tests under tests/ (notably in virt/upgrade/conftest.py and related test modules) and should be followed for any code paths that assume vm.vmi exists only when vm.ready is True.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-18T09:44:17.044Z
Learnt from: Anatw
Repo: RedHatQE/openshift-virtualization-tests PR: 3376
File: tests/network/general/test_ip_family_services.py:96-96
Timestamp: 2026-01-18T09:44:17.044Z
Learning: In the openshift-virtualization-tests repository, function-scoped fixtures must use pytest.fixture() with empty parentheses (not pytest.fixture without parentheses). This repo follows this convention despite Ruff PT001. Apply this consistently to all Python test files under tests/ (not just this one) to maintain repository-wide consistency.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-27T17:18:49.973Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 3619
File: tests/network/user_defined_network/test_user_defined_network.py:97-97
Timestamp: 2026-01-27T17:18:49.973Z
Learning: In tests that exercise lookup_iface_status_ip from libs.net.vmspec, rely on the function's built-in descriptive error messages for failures. Do not add extra assertion messages for IP presence checks using this function; instead, assert on the function behavior or catch its exceptions as appropriate. This reduces duplication and clarifies failures.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-02-23T16:31:34.505Z
Learnt from: vsibirsk
Repo: RedHatQE/openshift-virtualization-tests PR: 3883
File: utilities/unittests/test_os_utils.py:333-425
Timestamp: 2026-02-23T16:31:34.505Z
Learning: In integration/functional tests located under the tests/ directory, require assertion failure messages using the pattern: assert condition, "descriptive message". For unit tests under utilities/unittests/, rely on pytest's assertion introspection and descriptive test names; explicit failure messages are not required. This guidance helps maintain clear diagnostics for integration tests while keeping unit tests concise and leveraging pytest's built-in introspection.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-02-25T10:52:09.679Z
Learnt from: yossisegev
Repo: RedHatQE/openshift-virtualization-tests PR: 3873
File: tests/network/localnet/test_non_udn_localnet.py:7-9
Timestamp: 2026-02-25T10:52:09.679Z
Learning: In the RedHatQE/openshift-virtualization-tests repository, networking infrastructure requirements (nmstate, localnet bridge mappings, NIC availability) are not automatically tier3; they are considered standard test environment capabilities. Only tests with truly platform-specific, time-consuming, or bare-metal requirements should be marked as tier3. Apply this guidance to all Python tests under tests/, including tests/network/localnet/test_non_udn_localnet.py, ensuring tier3 designation is reserved for genuine platform-specific or complex scenarios rather than general networking infra necessities.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-03-19T10:36:59.023Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 4147
File: tests/network/upgrade/test_upgrade_network.py:166-177
Timestamp: 2026-03-19T10:36:59.023Z
Learning: In this repository’s pytest-based test files (under `tests/`), do not flag unused test method parameters/fixture arguments for removal when the parameters are intentionally kept only to enforce pytest fixture dependency ordering (e.g., an unused fixture like `bridge_on_one_node`). Treat this as an intentional convention consistent with other fixture definitions in the codebase, and do not open follow-up review issues for those unused parameters.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-03-25T11:24:07.687Z
Learnt from: jpeimer
Repo: RedHatQE/openshift-virtualization-tests PR: 4267
File: tests/storage/cross_cluster_live_migration/conftest.py:530-531
Timestamp: 2026-03-25T11:24:07.687Z
Learning: In this repo’s OpenShift virtualization tests, it is a standard pattern to call `to_dict()` on `ocp-resources` objects (e.g., `DataVolume`) without using its return value. The call is used only to populate the object’s `res` attribute, which is then read or mutated (e.g., `dv.res[...] = ...`). Do not flag this as an unused return value and do not request adding an inline comment just to justify it, since maintainers treat this behavior as consistent and intentional across the codebase.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-03-31T08:35:22.802Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 4318
File: tests/network/bandwidth/test_bandwidth.py:1-3
Timestamp: 2026-03-31T08:35:22.802Z
Learning: In this repository, when reviewing Python test modules under tests/, only require an STP link in the module docstring for STD (new-feature) tests. If the test module is a support-exception test tied to SUPPORTEX-* tickets (e.g., SUPPORTEX-29574) and therefore has no STP, do not flag missing STP links in the module docstring.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-03-31T10:39:33.409Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 4318
File: tests/network/l2_bridge/bandwidth/test_bandwidth.py:5-27
Timestamp: 2026-03-31T10:39:33.409Z
Learning: In this repository’s pytest suite, some STD (Standard Definition) placeholder tests under the `tests/` directory are intentionally implemented as docstring-only functions. They must be marked with `__test__ = False` immediately after the function definition (e.g., `test_x.__test__ = False`) to prevent pytest from collecting/executing them until the real test body exists. During code review, do not flag these as missing executable coverage and do not recommend removing or changing the `__test__ = False` setting when the placeholder pattern matches this accepted STD approach.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-04-02T09:03:57.004Z
Learnt from: ema-aka-young
Repo: RedHatQE/openshift-virtualization-tests PR: 4342
File: tests/storage/cdi_config/test_cdi_config.py:195-205
Timestamp: 2026-04-02T09:03:57.004Z
Learning: When calling `utilities.storage.get_dv_size_from_datasource()` in this repo, do not require an extra `None` guard for its return value if the `datasource` argument originates from a fixture that uses `ensure_exists=True` (e.g., `fedora_data_source_scope_module`). Those fixtures guarantee the datasource exists and provides a valid size, and other repo usages follow the same pattern without additional fail-fast/`None` checks; therefore, reviewers should not flag missing `None` guarding in this specific fixture-driven scenario.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-04-14T13:00:57.514Z
Learnt from: yossisegev
Repo: RedHatQE/openshift-virtualization-tests PR: 4328
File: tests/network/flat_overlay/test_multi_network_policy.py:32-32
Timestamp: 2026-04-14T13:00:57.514Z
Learning: In this repository’s pytest quarantine tests, do not flag or change `pytest.mark.jira(<id>, run=False)` as ineffective. `run=False` on `pytest.mark.jira()` is intentionally handled by the `pytest_jira` plugin, which conditionally skips the test when the referenced Jira issue is open. This is the correct behavior for **Category 1 (Product Bug)** quarantines; the `pytest.mark.xfail(run=False)` approach is reserved for **Category 2 (Automation Issue)** quarantines only (as documented in `docs/QUARANTINE_GUIDELINES.md`).
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-04-14T16:15:31.065Z
Learnt from: rnetser
Repo: RedHatQE/openshift-virtualization-tests PR: 4328
File: tests/network/flat_overlay/test_multi_network_policy.py:0-0
Timestamp: 2026-04-14T16:15:31.065Z
Learning: When reviewing Python tests, avoid redundant parentheses around f-strings in `pytest.mark.xfail` decorators. Prefer `pytest.mark.xfail(reason=f"...", run=False)` over `pytest.mark.xfail(reason=(f"..."), run=False)`, and flag/suggest removing the extra wrapping parentheses where found.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-04-14T16:15:33.012Z
Learnt from: rnetser
Repo: RedHatQE/openshift-virtualization-tests PR: 4328
File: tests/network/flat_overlay/test_multi_network_policy.py:32-32
Timestamp: 2026-04-14T16:15:33.012Z
Learning: For this repository (RedHatQE/openshift-virtualization-tests), when reviewing any PR with "Quarantine" in the title or a `quarantine` label, check compliance with `docs/QUARANTINE_GUIDELINES.md` in any affected pytest test code. Specifically: (1) For Category 1 (Product Bug), require `pytest.mark.jira("CNV-XXXXX", run=False)` and do not suggest replacing it with `xfail` (the `pytest_jira` plugin conditionally skips when the Jira issue is open). (2) For Category 2 (Automation Issue), require `pytest.mark.xfail(run=False, reason=...)` (pytest handles the skip). During review, flag quarantine PRs that use the wrong marker/category, omit a Jira ticket reference for Category 1, or use `run=False` in an incorrect context; raise these compliance questions even if later resolution confirms the marker was correct.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-04-21T19:08:39.771Z
Learnt from: servolkov
Repo: RedHatQE/openshift-virtualization-tests PR: 4542
File: tests/network/libs/bgp.py:333-354
Timestamp: 2026-04-21T19:08:39.771Z
Learning: In this codebase, when using `retry` imported from `timeout_sampler`, the decorated function must indicate success by returning a truthy value. The decorator retries until it encounters a truthy `sample` (i.e., it does `if sample: return sample`). Therefore, ensure `retry`-decorated functions include `return True` (or another truthy value) on success; avoid removing/altering it on the assumption that the caller ignores the return value. If the function returns `None` (or has no return statement), it will keep retrying and typically end in `TimeoutExpiredError`.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-04-26T11:44:20.150Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 4578
File: tests/network/localnet/nad_ref_change/test_nad_ref_change.py:18-18
Timestamp: 2026-04-26T11:44:20.150Z
Learning: In this repository’s STD (Standard Test Definition) PRs under `tests/**`, it’s expected that Polarion IDs may be placeholders (e.g., `pytest.mark.polarion("CNV-00000")`) while the test scenarios are still being agreed upon. If the test scenario is disabled for the moment (e.g., the scenario/module sets `__test__ = False`), reviewers should not treat placeholder Polarion IDs as a blocking issue. Real Polarion IDs should be filled in before or during the implementation PR.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-04-27T15:40:31.167Z
Learnt from: dalia-frank
Repo: RedHatQE/openshift-virtualization-tests PR: 4603
File: tests/data_protection/oadp/test_velero.py:101-130
Timestamp: 2026-04-27T15:40:31.167Z
Learning: When reviewing tests in RedHatQE/openshift-virtualization-tests, do not require STD-first workflow and STP/RFE/Jira-epic module docstring/traceability for tests that are being re-enabled after previously being blocked by a product bug (e.g., after a CNV Jira bug fix). Only enforce these STD-first and STP/RFE/Jira-epic module docstring requirements for genuinely new feature tests; previously blocked tests that are now unblocked/re-enabled should not be flagged as missing traceability.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-03T14:47:13.096Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 4569
File: tests/network/user_defined_network/rhel9_rhel10_cluster/test_connectivity.py:17-51
Timestamp: 2026-05-03T14:47:13.096Z
Learning: In this repository’s pytest suite (files under `tests/`), it is acceptable to have identically named `test_*` functions in different test modules (e.g., same function name in different `.../test_*.py` files). Do not request renaming solely to disambiguate function names across modules, since pytest node IDs remain unique and unambiguous due to the module path prefix. The team accepts the trade-off that `pytest -k <name>` will match tests in all modules containing that name.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-05T17:27:32.109Z
Learnt from: dshchedr
Repo: RedHatQE/openshift-virtualization-tests PR: 4739
File: tests/virt/node/descheduler/conftest.py:145-145
Timestamp: 2026-05-05T17:27:32.109Z
Learning: In this repo’s Python tests, `pyhelper_utils.shell.run_command` defaults to `check=True` (so it raises `subprocess.CalledProcessError` on non-zero exit). Therefore, don’t flag missing return-code handling/rc checks for bare calls like `run_command(...)` (or calls where `check` is not explicitly provided). Only flag cases where `check=False` is explicitly passed and the exit code/return value is ignored or not otherwise handled.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-05T18:28:01.097Z
Learnt from: dshchedr
Repo: RedHatQE/openshift-virtualization-tests PR: 4739
File: tests/virt/node/descheduler/conftest.py:142-146
Timestamp: 2026-05-05T18:28:01.097Z
Learning: In this repository, ignore Ruff rule PT022 in Python test files under `tests/`. If PT022 is triggered in a `pytest` fixture that uses `yield` but has no teardown code, treat it as an acceptable low-value nitpick and do not suggest changing the fixture to use `return` or otherwise flag the issue during code review.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-07T12:34:42.589Z
Learnt from: OhadRevah
Repo: RedHatQE/openshift-virtualization-tests PR: 4162
File: tests/install_upgrade_operators/crypto_policy/utils.py:320-323
Timestamp: 2026-05-07T12:34:42.589Z
Learning: When parsing the output of `openssl list -tls-groups`, remember it prints TLS group names on a single colon-separated line (e.g., `secp256r1:X25519:SecP256r1MLKEM768:...`). Parse individual TLS group names by splitting on `:` (e.g., `output.strip().split(':')`) rather than assuming one group per line. In code that specifically handles this `openssl list -tls-groups` output, do not flag `split(':')` as incorrect.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-08T12:31:26.895Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 4772
File: tests/network/l2_bridge/rhel9_rhel10_cluster/test_connectivity.py:20-22
Timestamp: 2026-05-08T12:31:26.895Z
Learning: In RedHatQE/openshift-virtualization-tests, any pytest test already marked with `pytest.mark.mixed_os_nodes` must not also be marked with `pytest.mark.tier3`. The `mixed_os_nodes` marker is responsible for both environment filtering and selecting the dedicated execution lane for dual-stream (mixed RHCOS 9 + RHCOS 10) cluster scenarios. Apply this rule to all test modules under `tests/` that use the `mixed_os_nodes` marker (e.g., `tests/network/l2_bridge/rhel9_rhel10_cluster/test_connectivity.py`).
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-05-12T18:18:20.607Z
Learnt from: Anatw
Repo: RedHatQE/openshift-virtualization-tests PR: 4833
File: tests/network/localnet/migration_stuntime/test_migration_stuntime.py:51-51
Timestamp: 2026-05-12T18:18:20.607Z
Learning: In RedHatQE/openshift-virtualization-tests, do not flag or suggest removing Python linter suppressions `# noqa: PID001` in files under `tests/`. `PID001` is a custom flake8 rule enforced by the Polarion ID checker plugin (not a standard Ruff rule). Therefore Ruff’s `RUF102` (“Invalid rule code in `# noqa`: PID001”) is a false positive in this repo. These `# noqa: PID001` comments are intentional and effective, commonly used on base-class `test_*` methods where the Polarion markers are defined on subclass overrides.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2025-12-22T16:27:40.244Z
Learnt from: yossisegev
Repo: RedHatQE/openshift-virtualization-tests PR: 3196
File: tests/network/upgrade/test_upgrade_network.py:4-4
Timestamp: 2025-12-22T16:27:40.244Z
Learning: For PRs that remove tests, rely on pytest --collect-only to verify the test discovery results (which tests are selected/deselected) and ensure the removal is clean and the test module remains functional. Full test execution is not required for test deletion PRs. This guideline applies to test files anywhere under the tests/ directory (e.g., tests/network/upgrade/test_upgrade_network.py) and should be used for similar test-deletion scenarios across the repository.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-07T09:52:12.342Z
Learnt from: yossisegev
Repo: RedHatQE/openshift-virtualization-tests PR: 3358
File: tests/network/sriov/test_sriov.py:21-21
Timestamp: 2026-01-07T09:52:12.342Z
Learning: When a PR only removes or modifies pytest markers in tests (e.g., removing pytest.mark.post_upgrade) and the test logic remains unchanged, prefer verifying with pytest --collect-only instead of running the full test suite. This validates that marker usage and test selection behavior are preserved. If the test logic changes, or markers affect behavior beyond collection, run the full test suite to confirm.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-18T13:18:48.808Z
Learnt from: EdDev
Repo: RedHatQE/openshift-virtualization-tests PR: 3273
File: tests/network/connectivity/test_ovs_linux_bridge.py:5-9
Timestamp: 2026-01-18T13:18:48.808Z
Learning: In tests/network/connectivity/test_ovs_linux_bridge.py and similar test files, prefer importing ipaddress as a module and using qualified calls like ipaddress.ip_interface(...) rather than from ipaddress import ip_interface. This preserves module context for readability, especially when chaining properties (e.g., ipaddress.ip_interface(...).ip). This is an intentional exception to the general rule favoring specific imports, and should apply to test files under the tests directory where module context aids understanding.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-18T14:51:50.846Z
Learnt from: yossisegev
Repo: RedHatQE/openshift-virtualization-tests PR: 3495
File: tests/network/third_part_ip_request/test_third_party_ip_request.py:4-12
Timestamp: 2026-01-18T14:51:50.846Z
Learning: In the openshift-virtualization-tests repository, tests consistently import pytest as a module (import pytest) and avoid from pytest import ...; this is the established pattern across 398+ test files. Do not flag or refactor imports to use specific pytest names in tests under tests/**. If a file already follows this pattern, leave it as is; this guideline applies broadly to Python test files under the tests directory.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-01-29T05:30:13.982Z
Learnt from: EdDev
Repo: RedHatQE/openshift-virtualization-tests PR: 3649
File: tests/network/user_defined_network/ip_specification/libipspec.py:1-4
Timestamp: 2026-01-29T05:30:13.982Z
Learning: In the openshift-virtualization-tests repository, Python imports should use module import style for the standard library 'json' (import json) rather than 'from json import ...'. This improves readability by making calls like json.loads and json.dumps explicit, and aligns with patterns used for modules like logging, ipaddress, cloudinit, and pytest. Apply this to all Python test files under tests/ (and similar test directories).
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-02-02T17:41:12.759Z
Learnt from: rnetser
Repo: RedHatQE/openshift-virtualization-tests PR: 3690
File: tests/after_cluster_deploy_sanity/test_after_cluster_deploy_sanity.py:65-65
Timestamp: 2026-02-02T17:41:12.759Z
Learning: In test files, keep test_* functions with simple one-line docstrings. For helper functions, utilities, and library code with non-obvious return values or side effects, use Google-style docstrings with Args, Returns, and Side effects sections. Do not require Google-style docstrings for pytest test functions themselves.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-02-03T07:34:34.184Z
Learnt from: RoniKishner
Repo: RedHatQE/openshift-virtualization-tests PR: 3697
File: tests/infrastructure/instance_types/test_common_vm_instancetype.py:53-98
Timestamp: 2026-02-03T07:34:34.184Z
Learning: In test files (Python, pytest), prefer using tier3 markers for categorization since tier2, tier1, and tier4 are not used in this repository. Do not rely on non-official markers; formalize and document the allowed markers in pytest.ini (or equivalent) to ensure consistent usage. When reviewing new tests, ensure markers align with this convention (only tier3, unless a project-wide decision defines additional tiers) and remove any unnecessary or undocumented markers.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-02-10T15:04:14.799Z
Learnt from: vsibirsk
Repo: RedHatQE/openshift-virtualization-tests PR: 3577
File: tests/virt/conftest.py:251-267
Timestamp: 2026-02-10T15:04:14.799Z
Learning: In Python tests, remove all bare time.sleep() calls. Replace with a waiting mechanism such as TimeoutSampler from the timeout_sampler package or a function decorated with retry (with appropriate timeout/conditions) to ensure determinism and avoid flakiness. This applies to all Python tests under the tests directory (not just this file) to maintain consistent waiting behavior across the suite.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-02-25T11:00:02.013Z
Learnt from: yossisegev
Repo: RedHatQE/openshift-virtualization-tests PR: 3873
File: tests/network/localnet/test_non_udn_localnet.py:19-27
Timestamp: 2026-02-25T11:00:02.013Z
Learning: In the test codebase, do not import from conftest.py files. This avoids import ambiguity in pytest. Do not import constants or helpers defined in conftest.py into tests. If a value is needed in both conftest.py and test files, duplicate it in both places or place it in a separate utility module that is importable by tests.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-03-29T13:51:25.599Z
Learnt from: jpeimer
Repo: RedHatQE/openshift-virtualization-tests PR: 4267
File: tests/storage/cross_cluster_live_migration/test_cclm.py:96-106
Timestamp: 2026-03-29T13:51:25.599Z
Learning: In this repository, follow the existing pytest convention for `pytest.mark.parametrize` argument names: use a single comma-separated string for `argnames` (e.g., `"dv_wait_timeout, vms_for_cclm"`), not a tuple (e.g., `("dv_wait_timeout", "vms_for_cclm")`). Do not flag or suggest changing `argnames` to a tuple. Also note that PT006 is not enforced by Ruff in this repo, so reviewers should not treat PT006 as a reason to alter the `argnames` format.
Applied to files:
tests/install_upgrade_operators/constants.pytests/virt/conftest.py
📚 Learning: 2026-03-16T18:23:45.130Z
Learnt from: albarker-rh
Repo: RedHatQE/openshift-virtualization-tests PR: 3580
File: tests/install_upgrade_operators/utils.py:337-364
Timestamp: 2026-03-16T18:23:45.130Z
Learning: In tests/install_upgrade_operators/utils.py (within the RedHatQE/openshift-virtualization-tests repo), assertion helper functions with self-descriptive names (e.g., assert_cnv_resource_container_env_image_not_in_upstream, assert_cnv_resource_container_image_not_in_upstream) do not require Google-style docstrings, even when they raise exceptions like ResourceMismatch. The function name itself should convey the assertion intent and the resource type being inspected. This guideline applies to similar tests in the same directory; if a function's purpose becomes ambiguous, consider adding a docstring or refactoring for clarity.
Applied to files:
tests/install_upgrade_operators/constants.py
📚 Learning: 2026-01-26T10:11:23.629Z
Learnt from: OhadRevah
Repo: RedHatQE/openshift-virtualization-tests PR: 2348
File: tests/observability/upgrade/conftest.py:1-7
Timestamp: 2026-01-26T10:11:23.629Z
Learning: In conftest.py files that define session-scoped fixtures which provision resources (e.g., VMs, namespaces) for upgrade testing, logging imports and structured INFO logs for setup are optional. Prefer quieter fixture setup to reduce log noise unless debugging is required. Apply this guideline across similar conftest.py fixtures that perform straightforward resource provisioning.
Applied to files:
tests/virt/conftest.py
📚 Learning: 2026-01-29T15:01:46.322Z
Learnt from: EdDev
Repo: RedHatQE/openshift-virtualization-tests PR: 3649
File: tests/network/user_defined_network/ip_specification/conftest.py:32-48
Timestamp: 2026-01-29T15:01:46.322Z
Learning: In test fixtures (e.g., tests/network/user_defined_network/ip_specification/conftest.py and other VM-related conftest.py files), after vm.start(), prefer calling vm.wait_for_agent_connected() as the validation step. Do not add vm.wait_for_ready_status(status=True) beforehand, since wait_for_agent_connected() ensures the guest OS is actually running (not just powered on) and provides a stronger, more reliable readiness check for test initialization.
Applied to files:
tests/virt/conftest.py
📚 Learning: 2026-03-03T12:01:13.888Z
Learnt from: jpeimer
Repo: RedHatQE/openshift-virtualization-tests PR: 4053
File: tests/storage/cross_cluster_live_migration/conftest.py:47-55
Timestamp: 2026-03-03T12:01:13.888Z
Learning: In test configuration files (for example tests/.../conftest.py), log the absence of optional CLI arguments or configurations at INFO level. Use WARNING only when missing configuration could cause a problem or be unusual. For parameters that may be present or absent by design, prefer INFO to reflect normality.
Applied to files:
tests/virt/conftest.py
📚 Learning: 2026-02-23T18:07:57.649Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 3930
File: tests/network/non_functional/ip_persistence/conftest.py:66-73
Timestamp: 2026-02-23T18:07:57.649Z
Learning: In all conftest.py files under the tests directory, do not flag or require fixes for Ruff FBT001 (Boolean-typed positional argument in function definition) in pytest fixture definitions. Pytest injects fixture parameters by name through dependency injection, so boolean positional arguments in fixture definitions are a false positive in this context. Apply this exception across all tests/* fixtures and document this behavior in review guidelines.
Applied to files:
tests/virt/conftest.py
📚 Learning: 2026-05-08T12:32:45.752Z
Learnt from: azhivovk
Repo: RedHatQE/openshift-virtualization-tests PR: 4772
File: tests/network/l2_bridge/rhel9_rhel10_cluster/conftest.py:26-26
Timestamp: 2026-05-08T12:32:45.752Z
Learning: In this repository, pytest fixture return annotations that use `collections.abc.Generator` with a single type parameter (e.g., `Generator[NetworkAttachmentDefinition]`) are the accepted shorthand. Do not flag `Generator[X]` as incorrect, and do not require the longer three-parameter form `Generator[X, None, None]` when used as a fixture return annotation. Apply this convention across all `conftest.py` files under `tests/`.
Applied to files:
tests/virt/conftest.py
📚 Learning: 2026-05-19T08:28:08.971Z
Learnt from: hmeir
Repo: RedHatQE/openshift-virtualization-tests PR: 4880
File: tests/install_upgrade_operators/deployment/conftest.py:25-28
Timestamp: 2026-05-19T08:28:08.971Z
Learning: In this repository, when reviewing targeted cherry-pick/backport PRs, do not fail the change on pytest fixture naming conventions if the fixture uses an action-verb name (e.g., `xfail_if_<condition>`) and its body calls `pytest.xfail(...)` as a side effect without returning or yielding a value. Treat this fixture pattern as acceptable for the backport until the deferred refactor to noun-based fixture names is done; for non-backport work, follow the usual fixture naming rules.
Applied to files:
tests/virt/conftest.py
📚 Learning: 2026-03-18T14:03:39.580Z
Learnt from: akri3i
Repo: RedHatQE/openshift-virtualization-tests PR: 3577
File: tests/virt/conftest.py:237-241
Timestamp: 2026-03-18T14:03:39.580Z
Learning: In tests/virt/conftest.py, the gpu_nodes_labeled_with_vgpu_config fixture should apply a single vGPU profile across all nodes in nodes_with_supported_gpus, derived from supported_gpu_device. This matches the test suite design that heterogeneous GPU models are not supported in this set, and all GPU nodes share the same device model. Ensure future changes preserve this assumption, and consider adding a brief comment or CI check to verify there are no heterogeneous GPUs among the nodes used by this fixture.
Applied to files:
tests/virt/conftest.py
📚 Learning: 2026-05-15T02:56:08.357Z
Learnt from: dshchedr
Repo: RedHatQE/openshift-virtualization-tests PR: 4871
File: tests/virt/node/descheduler/conftest.py:149-152
Timestamp: 2026-05-15T02:56:08.357Z
Learning: When reviewing Python test code under tests/virt/**, don’t flag missing “post-context” waits for node schedulability if the code is using the `node_mgmt_console` context manager from `utilities/virt.py`. Its `finally` teardown already uncordons the node (`oc adm uncordon`) and guarantees the schedulability wait via `wait_for_node_schedulable_status(node=..., status=True)` (and then calls `wait_for_kv_stabilize()`). Only require additional schedulability waiting when `node_mgmt_console` is not in use for that node-management flow.
Applied to files:
tests/virt/conftest.py
🪛 Ruff (0.15.13)
tests/virt/conftest.py
[warning] 261-261: Unused function argument: hco_with_disable_mdev_configuration
(ARG001)
🔀 Multi-repo context RedHatQE/openshift-virtualization-tests-design-docs
Findings for RedHatQE/openshift-virtualization-tests-design-docs
- Multiple STP documents reference HCO feature gates (context that HCO feature-gate flags are part of testing expectations), e.g.:
- stps/sig-storage/cbt.md lines mentioning "HCO feature gate for CBT enabled" and related environment notes (matches at lines ~96,100,110,118,134,142,169) [::RedHatQE/openshift-virtualization-tests-design-docs::stps/sig-storage/cbt.md]
- stps/sig-virt/dual-stream-cluster-rhcos9-rhcos10/network.md references HCO and OCP 4.22 support and feature-gate context (multiple matches) [::RedHatQE/openshift-virtualization-tests-design-docs::stps/sig-virt/dual-stream-cluster-rhcos9-rhcos10/network.md]
- stps/sig-iuo/multiarch_arm_support.md references HCO multi-arch support (match) [::RedHatQE/openshift-virtualization-tests-design-docs::stps/sig-iuo/multiarch_arm_support.md]
Assessment
- No code consumers or technical references to the new DISABLE_MDEV_CONFIGURATION constant were found in this repo; only documentation/test-plan references to HCO feature gates in general. These docs provide contextual relevance but do not indicate direct API or constant usage that would break.
🔇 Additional comments (3)
utilities/constants.py (1)
409-409: LGTM!tests/install_upgrade_operators/constants.py (1)
1-1: LGTM!tests/virt/conftest.py (1)
10-10: LGTM!Also applies to: 44-45, 51-51
Contributor
|
Clean rebase detected — no code changes compared to previous head ( |
openshift-virtualization-qe-bot
added
tox:verify-tc-requirement-polarion:running
Contributor
|
✅ Action performedFull review finished. |
Contributor
Author
|
/retest all |
|
Clean rebase detected — no code changes compared to previous head ( |
vsibirsk
requested changes
Jun 24, 2026
Contributor
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@tests/virt/node/gpu/utils.py`:
- Around line 148-149: The teardown in the vGPU helper is too broad because it
clears labels from nodes tracked by nodes_with_supported_gpus, which can overlap
between the package-scoped and session-scoped fixture chains. Update the cleanup
in the shared fixture helper to only remove nvidia.com/vgpu.config.state from
the exact nodes labeled by that fixture instance, using the same nodes
collection established in the corresponding setup path, so one fixture’s
teardown does not affect another still-active fixture.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: ASSERTIVE
Plan: Pro
Run ID: e312f1df-6cae-40fe-b212-225933b38711
📒 Files selected for processing (9)
tests/install_upgrade_operators/constants.pytests/virt/conftest.pytests/virt/node/gpu/utils.pytests/virt/node/gpu/vgpu/conftest.pytests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.pytests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.pytests/virt/upgrade_custom/vgpu/conftest.pytests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.pyutilities/constants/hco.py
hmeir
approved these changes
Jun 24, 2026
Contributor
Author
|
@coderabbitai |
Contributor
|
Why: a full review reruns analysis across the entire PR, which helps catch issues introduced by rebases, overlapping merges, or fixture dependency changes that may not show up in a partial pass. ✅ Action performedFull review finished. |
Contributor
Author
|
/build-and-push-container |
|
New container for quay.io/openshift-cnv/openshift-virtualization-tests:pr-4959 published |
Contributor
|
Clean rebase detected — no code changes compared to previous head ( |
Contributor
|
Clean rebase detected — no code changes compared to previous head ( |
Contributor
|
/retest all Auto-triggered: Files in this PR were modified by merged PR #5318. Overlapping filestests/virt/conftest.py |
dshchedr
previously approved these changes
Jun 30, 2026
Add hco_with_disable_mdev_configuration fixture that enables the disableMDevConfiguration feature gate in HCO before any vGPU node labeling occurs. This ensures CNV does not configure mediated devices, as vGPU configuration is handled by NVIDIA GPU Operator. Changes: - Add DISABLE_MDEV_CONFIGURATION constant to utilities/constants.py - Create hco_with_disable_mdev_configuration session-scoped fixture - Give each second vGPU test VM its own root disk instead of sharing the class-scoped data volume template, which caused XFS corruption and guest agent failures when VMs landed on different nodes. Signed-off-by: akri3i <guptaakriti70@gmail.com>
Collaborator
|
/approve |
Contributor
Author
|
/verified |
Contributor
Test execution plan request detailsCRITICAL: You MUST post an inline review comment on the first changed line of the first file. As an expert software testing engineer, analyze all modified files in this PR and create a targeted test execution plan. Analysis Requirements:
Output rules: Your deliverable: Test Execution Plan
Use these formats:
Real test commands (MANDATORY when changes affect session/runtime code): When the affected code runs at session/collection time (conftest fixtures, pytest plugins,
Example output for a session-startup fix: Guidelines:
Hardware-Related Checks (SR-IOV, GPU, DPDK): When PR modifies fixtures for hardware-specific resources:
CRITICAL WORKFLOW COMPLETION RULES: When responding to this test execution plan request, you MUST follow these rules EXACTLY:
ABSOLUTE PROHIBITIONS (violating these creates empty/meaningless reviews):
Remember: The pull request review is visible to users. Additional comments are redundant noise. CRITICAL — Verification rules for smoke and gating decisions: |
Contributor
|
Tip For best results, initiate chat on the files or code changes. 🧩 Analysis chain🏁 Script executed: #!/bin/bash
set -euo pipefail
repo="RedHatQE/openshift-virtualization-tests"
pr="4959"
# Ensure we have the exact PR metadata and diff context.
pr_json="$(gh api "repos/$repo/pulls/$pr")"
base_sha="$(jq -r '.base.sha' <<<"$pr_json")"
head_sha="$(jq -r '.head.sha' <<<"$pr_json")"
echo "PR base: $base_sha"
echo "PR head: $head_sha"
# Fetch the exact SHAs for local diffing if needed.
git fetch --no-tags --depth=1 origin "$base_sha" "$head_sha" >/dev/null 2>&1 || true
echo "=== Changed files (ordered) ==="
gh api "repos/$repo/pulls/$pr/files?per_page=100" --jq '.[] | .filename'
echo "=== First changed file and first changed line ==="
python - <<'PY'
import json, subprocess, re, sys
repo = "RedHatQE/openshift-virtualization-tests"
pr = "4959"
files = json.loads(subprocess.check_output(
["gh", "api", f"repos/{repo}/pulls/{pr}/files?per_page=100"],
text=True,
))
first = files[0]
patch = first.get("patch") or ""
path = first["filename"]
first_right_line = None
new_line = None
for line in patch.splitlines():
if line.startswith("@@"):
m = re.search(r"\+(\d+)(?:,(\d+))?", line)
if m:
new_line = int(m.group(1))
continue
if new_line is None:
continue
if line.startswith("+") and not line.startswith("+++"):
first_right_line = new_line
break
if not line.startswith("-"):
new_line += 1
print(json.dumps({
"path": path,
"first_right_line": first_right_line,
"patch": patch,
}, indent=2))
PY
echo "=== AST outlines for key changed files ==="
for f in \
tests/virt/node/gpu/utils.py \
tests/virt/node/gpu/vgpu/conftest.py \
tests/virt/upgrade_custom/vgpu/conftest.py \
tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py \
tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py \
tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py \
utilities/constants/hco.py \
tests/install_upgrade_operators/constants.py
do
if [ -f "$f" ]; then
echo "--- $f ---"
ast-grep outline "$f" --view expanded || true
fi
done
echo "=== Modified symbols in utilities/libs/shared helper files ==="
echo "--- git diff for utilities/constants/hco.py ---"
git diff --unified=0 "$base_sha...$head_sha" -- utilities/constants/hco.py || true
echo "--- git diff for tests/virt/node/gpu/utils.py ---"
git diff --unified=0 "$base_sha...$head_sha" -- tests/virt/node/gpu/utils.py || true
echo "=== Direct and transitive callers of changed symbols / fixtures ==="
rg -n -C2 '\bDISABLE_MDEV_CONFIGURATION\b' tests/ utilities/ --type=py || true
rg -n -C2 '\bHCO_DEFAULT_FEATUREGATES\b' tests/ --type=py || true
rg -n -C2 '\bvgpu_node_labels_applied\s*\(' tests/ --type=py || true
rg -n -C2 '\bwait_for_nvidia_vgpu_manager\s*\(' tests/ --type=py || true
rg -n -C2 '\bverify_mdev_bus_available\s*\(' tests/ --type=py || true
rg -n -C2 '\bgpu_nodes_labeled_with_vgpu_config(_session_scope)?\b' tests/ --type=py || true
rg -n -C2 '\bvgpu_ready_nodes(_session_scope)?\b' tests/ --type=py || true
rg -n -C2 '\bnon_existent_mdev_bus_nodes(_session_scope)?\b' tests/ --type=py || true
echo "=== Smoke-marked files ==="
smoke_files="$(rg -l '`@pytest`\.mark\.smoke' tests/ --type=py || true)"
printf '%s\n' "$smoke_files"
echo "=== Gating-marked files ==="
gating_files="$(rg -l '`@pytest`\.mark\.gating' tests/ --type=py || true)"
printf '%s\n' "$gating_files"
echo "=== Smoke/gating markers inside affected areas ==="
printf '%s\n' "$smoke_files" | rg 'tests/virt/node/gpu/vgpu|tests/virt/upgrade_custom/vgpu|tests/install_upgrade_operators' || true
printf '%s\n' "$gating_files" | rg 'tests/virt/node/gpu/vgpu|tests/virt/upgrade_custom/vgpu|tests/install_upgrade_operators' || true
echo "=== Specific marker check for likely affected tests ==="
for f in \
tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py \
tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py \
tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py \
tests/install_upgrade_operators/feature_gates/test_update_featuregate_hco.py \
tests/install_upgrade_operators/feature_gates/test_default_featuregates.py
do
echo "--- $f ---"
rg -n -C2 '`@pytest`\.mark\.(smoke|gating)' "$f" || true
done
echo "=== Test definitions for lightweight command selection ==="
for f in \
tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py \
tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py \
tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py \
tests/install_upgrade_operators/feature_gates/test_update_featuregate_hco.py \
tests/install_upgrade_operators/feature_gates/test_default_featuregates.py
do
echo "--- $f ---"
rg -n '^(class |def test_)' "$f" || true
done
echo "=== Relevant conftest dependency slices ==="
sed -n '1,240p' tests/virt/node/gpu/vgpu/conftest.py || true
sed -n '1,260p' tests/virt/upgrade_custom/vgpu/conftest.py || true
sed -n '1,240p' tests/virt/node/gpu/utils.py || true
sed -n '1,200p' tests/install_upgrade_operators/constants.py || true
sed -n '1,120p' utilities/constants/hco.py || true
echo "=== Build and post inline test execution plan comment ==="
python - <<'PY'
import json, re, subprocess, textwrap, sys
repo = "RedHatQE/openshift-virtualization-tests"
pr = "4959"
files = json.loads(subprocess.check_output(
["gh", "api", f"repos/{repo}/pulls/{pr}/files?per_page=100"],
text=True,
))
first = files[0]
path = first["filename"]
patch = first.get("patch") or ""
first_right_line = None
new_line = None
for line in patch.splitlines():
if line.startswith("@@"):
m = re.search(r"\+(\d+)(?:,(\d+))?", line)
if m:
new_line = int(m.group(1))
continue
if new_line is None:
continue
if line.startswith("+") and not line.startswith("+++"):
first_right_line = new_line
break
if not line.startswith("-"):
new_line += 1
if first_right_line is None:
raise SystemExit("Could not determine first changed line on RHS")
body = textwrap.dedent("""\
## Test Execution Plan
**Run smoke tests: False**
Verified with `rg -l '`@pytest.mark.smoke`' tests/`: no smoke-marked file in the affected set depends on `tests/virt/node/gpu/utils.py::{vgpu_node_labels_applied, wait_for_nvidia_vgpu_manager, verify_mdev_bus_available}`, the new vGPU package/session fixtures, or `utilities/constants/hco.py::DISABLE_MDEV_CONFIGURATION`.
**Run gating tests: False**
Verified with `rg -l '`@pytest.mark.gating`' tests/`: no gating-marked file is in the affected call/fixture chain. The direct callers of `DISABLE_MDEV_CONFIGURATION` / `HCO_DEFAULT_FEATUREGATES` and the vGPU fixture consumers affected by this PR are not gating-marked.
**Affected tests to run**
- `tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py`
- `tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py`
- `tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py`
- `tests/install_upgrade_operators/feature_gates/test_update_featuregate_hco.py`
- `tests/install_upgrade_operators/feature_gates/test_default_featuregates.py`
**Why these tests**
- `tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py` / `tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py` depend on `non_existent_mdev_bus_nodes -> vgpu_ready_nodes -> gpu_nodes_labeled_with_vm_vgpu -> gpu_nodes_labeled_with_vgpu_config -> hco_with_disable_mdev_configuration`, and their VM fixtures now build per-test data volume templates from `golden_image_data_source_for_test_scope_class`.
- `tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py` now uses the renamed session-scoped chain `non_existent_mdev_bus_nodes_session_scope -> vgpu_ready_nodes_session_scope -> ... -> hco_with_disable_mdev_configuration_session_scope`.
- `tests/install_upgrade_operators/feature_gates/test_update_featuregate_hco.py` and `tests/install_upgrade_operators/feature_gates/test_default_featuregates.py` are the direct feature-gate callers impacted by moving `DISABLE_MDEV_CONFIGURATION` to `utilities/constants/hco.py`.
**Real tests (cluster required)**
Error path (the fix: vGPU tests that previously reused a shared template / PVC):
`pytest tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py -vv`
Expected: vGPU VMs start with unique data volume templates/root disks; no RWX PVC reuse-induced XFS corruption or guest-agent failures when scheduled on different nodes.
Happy path (regression: HCO gate + session/runtime fixture chain still works):
`pytest tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py -vv`
Expected: session-scoped HCO reconcile, node labeling, NVIDIA vGPU manager readiness, and `mdev_bus` verification complete successfully.
Collection safety on clusters **without** NVIDIA GPU hardware:
`pytest tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py --collect-only`
Expected: collection succeeds cleanly without fixture startup failures when GPU hardware is absent.
""")
# Post as a single inline non-blocking review comment.
head_sha = json.loads(subprocess.check_output(
["gh", "api", f"repos/{repo}/pulls/{pr}"],
text=True,
))["head"]["sha"]
cmd = [
"gh", "api", f"repos/{repo}/pulls/{pr}/reviews",
"--method", "POST",
"--field", f"commit_id={head_sha}",
"--field", "event=COMMENT",
"--field", "body=",
"--field", f"comments[][path]={path}",
"--field", f"comments[][line]={first_right_line}",
"--field", "comments[][side]=RIGHT",
"--field", f"comments[][body]={body}",
]
subprocess.run(cmd, check=True)
PYLength of output: 71468 🏁 Script executed: #!/bin/bash
set -euo pipefail
repo="RedHatQE/openshift-virtualization-tests"
pr="4959"
echo "=== Inspect install_upgrade_operators conftest impact ==="
ast-grep outline tests/install_upgrade_operators/conftest.py --view expanded || true
rg -n -C3 '\bHCO_DEFAULT_FEATUREGATES\b|\bDISABLE_MDEV_CONFIGURATION\b|`@pytest`\.fixture|autouse=True|scope="session"|scope="package"' \
tests/install_upgrade_operators/conftest.py tests/install_upgrade_operators/feature_gates/test_default_featuregates.py \
tests/install_upgrade_operators/feature_gates/test_update_featuregate_hco.py \
tests/install_upgrade_operators/deployment/test_hco_deployment_params.py \
tests/install_upgrade_operators/product_upgrade/test_upgrade.py \
tests/install_upgrade_operators/hco_enablement_golden_image_updates/test_custom_golden_images_namespace.py \
--type=py || true
echo "=== Build final plan body and retry posting ==="
python - <<'PY'
import json
import re
import subprocess
import textwrap
repo = "RedHatQE/openshift-virtualization-tests"
pr = "4959"
files = json.loads(subprocess.check_output(
["gh", "api", f"repos/{repo}/pulls/{pr}/files?per_page=100"],
text=True,
))
first = files[0]
path = first["filename"]
patch = first.get("patch") or ""
first_right_line = None
new_line = None
for line in patch.splitlines():
if line.startswith("@@"):
m = re.search(r"\+(\d+)(?:,(\d+))?", line)
if m:
new_line = int(m.group(1))
continue
if new_line is None:
continue
if line.startswith("+") and not line.startswith("+++"):
first_right_line = new_line
break
if not line.startswith("-"):
new_line += 1
if first_right_line is None:
raise SystemExit("Could not determine first changed RHS line")
body = textwrap.dedent("""\
## Test Execution Plan
**Run smoke tests: False**
Verified with `rg -l '`@pytest.mark.smoke`' tests/`: none of the smoke-marked files depend on the changed vGPU fixture chain or the imported `utilities/constants/hco.py::DISABLE_MDEV_CONFIGURATION` constant.
**Run gating tests: False**
Verified with `rg -l '`@pytest.mark.gating`' tests/`: no gating-marked test is in the affected call/fixture chain. The impacted vGPU consumers and the direct feature-gate tests are not gating-marked, and the `tests/install_upgrade_operators/constants.py` change is an import-only refactor with unchanged value.
**Affected tests to run**
- `tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py`
- `tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py`
- `tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py`
- `tests/install_upgrade_operators/feature_gates/test_update_featuregate_hco.py`
- `tests/install_upgrade_operators/feature_gates/test_default_featuregates.py`
**Why these tests**
- `tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py` and `tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py` use `non_existent_mdev_bus_nodes -> vgpu_ready_nodes -> gpu_nodes_labeled_with_vm_vgpu -> gpu_nodes_labeled_with_vgpu_config -> hco_with_disable_mdev_configuration`, and this PR also changes their second-vGPU VM fixtures to generate unique data volume templates from `golden_image_data_source_for_test_scope_class`.
- `tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py` uses the renamed session-scoped chain `non_existent_mdev_bus_nodes_session_scope -> vgpu_ready_nodes_session_scope -> gpu_nodes_labeled_with_vm_vgpu_session_scope -> gpu_nodes_labeled_with_vgpu_config_session_scope -> hco_with_disable_mdev_configuration_session_scope`.
- `tests/install_upgrade_operators/feature_gates/test_update_featuregate_hco.py` and `tests/install_upgrade_operators/feature_gates/test_default_featuregates.py` are the direct callers of the moved `DISABLE_MDEV_CONFIGURATION` / `HCO_DEFAULT_FEATUREGATES` constants.
**Real tests (cluster required)**
Error path (the fix: avoid shared-template/PVC reuse across second-vGPU VMs):
`pytest tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py -vv`
Expected: each VM gets its own DV/root disk template; no RWX PVC reuse-induced XFS corruption or guest-agent failures when VMs land on different nodes.
Happy path (regression: HCO gate + runtime fixture chain still works):
`pytest tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py -vv`
Expected: session-scoped HCO reconcile, node labeling, NVIDIA vGPU manager readiness, and `mdev_bus` verification complete successfully.
Collection safety on clusters **without** NVIDIA GPU hardware:
`pytest tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py --collect-only`
Expected: collection succeeds without fixture startup failures when GPU hardware is absent.
""")
head_sha = json.loads(subprocess.check_output(
["gh", "api", f"repos/{repo}/pulls/{pr}"],
text=True,
))["head"]["sha"]
def run(cmd):
return subprocess.run(cmd, text=True, capture_output=True)
# Retry inline comment using the direct review-comment endpoint.
inline_cmd = [
"gh", "api", f"repos/{repo}/pulls/{pr}/comments",
"--method", "POST",
"--field", f"body={body}",
"--field", f"commit_id={head_sha}",
"--field", f"path={path}",
"--field", f"line={first_right_line}",
"--field", "side=RIGHT",
]
res = run(inline_cmd)
print("INLINE_STATUS", res.returncode)
if res.stdout:
print(res.stdout)
if res.stderr:
print(res.stderr)
if res.returncode == 0:
raise SystemExit(0)
# Fallback: regular PR comment, as explicitly allowed by the bot instructions.
fallback_body = body + "\n\n_Inline comment fallback: GitHub API returned `Resource not accessible by integration` for review-comment creation._"
issue_cmd = [
"gh", "api", f"repos/{repo}/issues/{pr}/comments",
"--method", "POST",
"--field", f"body={fallback_body}",
]
res2 = run(issue_cmd)
print("FALLBACK_STATUS", res2.returncode)
if res2.stdout:
print(res2.stdout)
if res2.stderr:
print(res2.stderr)
if res2.returncode != 0:
raise SystemExit(res2.returncode)
PYLength of output: 29335
CRITICAL: I could not complete the required post because GitHub rejected both the inline review comment API and the fallback PR comment API with Why this matters: the workflow requires posting on Files Changed to avoid noisy/empty reviews, but that also means missing write permission is a hard blocker. Target location
Exact Test Execution Plan to post ## Test Execution Plan
**Run smoke tests: False**
Verified with `rg -l '`@pytest.mark.smoke`' tests/`: none of the smoke-marked files depend on the changed vGPU fixture chain or the imported `utilities/constants/hco.py::DISABLE_MDEV_CONFIGURATION` constant.
**Run gating tests: False**
Verified with `rg -l '`@pytest.mark.gating`' tests/`: no gating-marked test is in the affected call/fixture chain. The impacted vGPU consumers and the direct feature-gate tests are not gating-marked, and the `tests/install_upgrade_operators/constants.py` change is an import-only refactor with unchanged value.
**Affected tests to run**
- `tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py`
- `tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py`
- `tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py`
- `tests/install_upgrade_operators/feature_gates/test_update_featuregate_hco.py`
- `tests/install_upgrade_operators/feature_gates/test_default_featuregates.py`
**Why these tests**
- `tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py` and `tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py` use `non_existent_mdev_bus_nodes -> vgpu_ready_nodes -> gpu_nodes_labeled_with_vm_vgpu -> gpu_nodes_labeled_with_vgpu_config -> hco_with_disable_mdev_configuration`, and this PR also changes their second-vGPU VM fixtures to generate unique data volume templates from `golden_image_data_source_for_test_scope_class`.
- `tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py` uses the renamed session-scoped chain `non_existent_mdev_bus_nodes_session_scope -> vgpu_ready_nodes_session_scope -> gpu_nodes_labeled_with_vm_vgpu_session_scope -> gpu_nodes_labeled_with_vgpu_config_session_scope -> hco_with_disable_mdev_configuration_session_scope`.
- `tests/install_upgrade_operators/feature_gates/test_update_featuregate_hco.py` and `tests/install_upgrade_operators/feature_gates/test_default_featuregates.py` are the direct callers of the moved `DISABLE_MDEV_CONFIGURATION` / `HCO_DEFAULT_FEATUREGATES` constants.
**Real tests (cluster required)**
Error path (the fix: avoid shared-template/PVC reuse across second-vGPU VMs):
`pytest tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py -vv`
Expected: each VM gets its own DV/root disk template; no RWX PVC reuse-induced XFS corruption or guest-agent failures when VMs land on different nodes.
Happy path (regression: HCO gate + runtime fixture chain still works):
`pytest tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py -vv`
Expected: session-scoped HCO reconcile, node labeling, NVIDIA vGPU manager readiness, and `mdev_bus` verification complete successfully.
Collection safety on clusters **without** NVIDIA GPU hardware:
`pytest tests/virt/node/gpu/vgpu/test_rhel_vm_with_vgpu.py tests/virt/node/gpu/vgpu/test_windows_vm_with_vgpu.py tests/virt/upgrade_custom/vgpu/test_vgpu_vm_upgrade.py --collect-only`
Expected: collection succeeds without fixture startup failures when GPU hardware is absent.If write permissions are restored for this integration, I can retry the required inline post. |
vsibirsk
approved these changes
Jul 2, 2026
Contributor
|
Successfully removed PR tag: quay.io/openshift-cnv/openshift-virtualization-tests:pr-4959. |
Contributor
|
New container for quay.io/openshift-cnv/openshift-virtualization-tests:latest published |
Contributor
Author
|
/cherry-pick cnv-4.22 |
1 similar comment
Contributor
Author
|
/cherry-pick cnv-4.22 |
|
Manual cherry-pick is needed |
|
Cherry-pick label already present for: |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
11 participants
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.
Add hco_with_disable_mdev_configuration fixture that enables the disableMDevConfiguration feature gate in HCO before any vGPU node labeling occurs. This ensures CNV does not configure mediated devices, as vGPU configuration is handled by NVIDIA GPU Operator. Changes:
What this PR does / why we need it:
This PR fixes vGPU test setup for clusters where NVIDIA GPU Operator owns vGPU configuration.
Feature gate: Adds a hco_with_disable_mdev_configuration fixture that enables disableMDevConfiguration in HCO before any vGPU node labeling. Without this, CNV auto-configures mediated devices and conflicts with GPU Operator’s vGPU setup.
Separate root disks: Second-VM fixtures (gpu_vmb, node_mdevtype_gpu_vm, gpu_vmc) now build a unique data volume template per VM instead of reusing the class-scoped template. The shared template gave both VMs the same RWX PVC name; when they landed on different nodes, vm2 hit XFS corruption and never reached AgentConnected=True (e.g. TestNodeMDEVTypeVGPURHELGPUSSpec).
Which issue(s) this PR fixes:
Special notes for reviewer:
jira-ticket:
Summary by CodeRabbit
mdev_busis available on targeted nodes before tests proceed.disableMdevConfigurationfeature gate during setup.mdev_busabsence validation is applied.