Skip to content

QA fixups - #2664

Merged
kmcdonell merged 10 commits into
performancecopilot:mainfrom
kmcdonell:wip
Jul 25, 2026
Merged

QA fixups#2664
kmcdonell merged 10 commits into
performancecopilot:mainfrom
kmcdonell:wip

Conversation

@kmcdonell

Copy link
Copy Markdown
Member

Title says it all.

Left over from a long-forgotten debugging session.  Triggered a
remake of 272.out to match.
localhost -> local: change in src/chknumval.c made one diagnostic
slightly different for qa/273.
…in pmcd PMDA

Needed to change
    sudo pmstore -h '$LOCALHOST_FULL' pmcd.control.debug ...
which will not work because it is run on a remote host aimed at
pmcd on the host where the QA test is being run and the access
controls bite early and break the test.

Replace this with
    pmstore -h '$LOCALHOST_FULL' sample.write_me 42
which will work because the sample PMDA does not have the same
store access controls as the pmcd PMDA.

Actually setting pmcd.control.debug or not has no functional impact
on this QA test, we're just after *some* pmstore operation.
The comment in the test explains:
    Based on smack-a-mole testing, it would appear that this PMDA is
    only going to work in an AWS context where there is at least one
    RDS endpoint.

This moves qa/1998 from failing+triaged to notrun *everywhere* in CI
and the QA Farm.

We really need the developers of the rds PMDA (Oracle) to step up to
the plate here and help us provision some QA environment that would
allow this PMDA to be tested.
Change chknumval.c to use "mmv.control.reload" instead of
"pmcd.control.sighup" (in the default case) to dodge new access
controls for pmStore in the pmcd PMDA.

Triggered remake of qa/043.
Changed access controls for pmStore and the pmcd PMDA meant
src/agenttimeout.c needed to create a pmNewContext using "local:"
rather than "localhost".

But this change needed to be reflected in qa/023 in the [access]
clause and the filtering (connection is now via a Unix domain socket,
not a TCP/IP port), which in turn trigeers remakes for 023.out.1 and
023.out.2.
May fix some transient failures in the QA Farm.
@coderabbitai

coderabbitai Bot commented Jul 25, 2026

Copy link
Copy Markdown
Contributor

Review Change Stack

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: Repository YAML (base), Repository UI (inherited), Organization UI (inherited)

Review profile: CHILL

Plan: Pro Plus

Run ID: 4db33c39-1fb4-4d40-a7ad-131c725d658c

📥 Commits

Reviewing files that changed from the base of the PR and between 4299c32 and 9c8434d.

📒 Files selected for processing (1)
  • qa/1998
🚧 Files skipped from review as they are similar to previous changes (1)
  • qa/1998

📝 Walkthrough

Summary by CodeRabbit

  • Tests
    • Improved QA determinism with clearer synchronization and more reliable PMCD timeout validation.
    • Updated local connection/access checks to reflect current transport behavior and client identifier reporting.
    • Strengthened log service restart handling with stricter failure detection and readiness waits.
    • Added a cloud PMDA pre-flight gate (requires available tooling and at least one endpoint) before running the RDS test.
    • Removed an obsolete privileged pmstore action and updated metric validation coverage.
  • Chores
    • Removed an outdated triage exception for the RDS monitoring test.

Walkthrough

QA tests and helper validation were updated for PMCD local transport normalization, timeout synchronization, pmlogger recovery, store access checks, RDS environment gating, and the default metric list.

Changes

QA maintenance

Layer / File(s) Summary
PMCD timeout normalization
qa/023, qa/023.out.*
PMCD socket identifiers are normalized consistently, local transport access is expected, and timeout execution includes a triage wait point.
Service and store checks
qa/032, qa/069, qa/272
Cleanup restores pmlogger, service control validates captured output and wait conditions, and store access commands are adjusted or removed.
Environment gating
qa/1998, qa/triaged
The RDS test checks for the AWS CLI and an available endpoint before running, and its obsolete triage entry is removed.
Default metric validation
qa/src/chknumval.c
The default metric changes from pmcd.control.sighup to mmv.control.reload.

Possibly related PRs

Poem

A rabbit checks the socket’s trail,
Waits for timeouts without fail.
Pmlogger wakes, the stores align,
RDS gates keep tests in line.
New metrics hop through QA’s door.

🚥 Pre-merge checks | ✅ 2 | ❌ 2

❌ Failed checks (2 inconclusive)

Check name Status Explanation Resolution
Title check ❓ Inconclusive The title is related to the PR, but it is too generic to identify the main QA changes. Use a concise title that names the main QA fix or affected test instead of a broad label.
Description check ❓ Inconclusive The description is related to the PR, but it is only a placeholder and conveys no useful details. Replace it with a brief summary of the QA fixes or tests changed.
✅ Passed checks (2 passed)
Check name Status Explanation
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.

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.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@coderabbitai coderabbitai Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Actionable comments posted: 2

🤖 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 `@qa/032`:
- Around line 27-29: Update the pmlogger restart flow in qa/032 to capture and
validate _service pmlogger restart’s status before passing its output to
_filter_pcp_start. Preserve the existing failure path through _exit 1 and ensure
filtering cannot mask a failed restart in this /bin/sh script.

In `@qa/1998`:
- Around line 40-42: Update the RDS endpoint check following the count
assignment to invoke _notrun with || when the test condition is true. Preserve
the existing “no RDS endpoints found” message and ensure the test exits through
the skip helper instead of treating _notrun as an argument to [.
🪄 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: Repository YAML (base), Repository UI (inherited), Organization UI (inherited)

Review profile: CHILL

Plan: Pro Plus

Run ID: 86bd123a-3494-4a26-bab3-9cc1ad081294

📥 Commits

Reviewing files that changed from the base of the PR and between 74a4a4f and 21c8b49.

⛔ Files ignored due to path filters (3)
  • qa/043.out is excluded by !**/*.out
  • qa/272.out is excluded by !**/*.out
  • qa/273.out is excluded by !**/*.out
📒 Files selected for processing (9)
  • qa/023
  • qa/023.out.1
  • qa/023.out.2
  • qa/032
  • qa/069
  • qa/1998
  • qa/272
  • qa/src/chknumval.c
  • qa/triaged
💤 Files with no reviewable changes (2)
  • qa/272
  • qa/triaged

Comment thread qa/032 Outdated
Comment thread qa/1998 Outdated

@coderabbitai coderabbitai Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

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 `@qa/1998`:
- Line 42: Reverse the preflight logic in the RDS endpoint check: update the
condition around count so empty or zero count invokes _notrun, while a positive
endpoint count allows the test to continue. Replace the current OR-based guard
with the appropriate AND-based behavior without changing the surrounding test
flow.
🪄 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: Repository YAML (base), Repository UI (inherited), Organization UI (inherited)

Review profile: CHILL

Plan: Pro Plus

Run ID: a910d83e-63f4-4b50-a838-a2445a49fa6e

📥 Commits

Reviewing files that changed from the base of the PR and between 21c8b49 and 4299c32.

📒 Files selected for processing (2)
  • qa/032
  • qa/1998

Comment thread qa/1998 Outdated
@kmcdonell
kmcdonell merged commit 0c6ddfc into performancecopilot:main Jul 25, 2026
0 of 16 checks passed
This was referenced Jul 26, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant