Skip to content

Expect 48 files in pmm-admin summary zip (sep-provision.log) - #1166

Draft
claude[bot] wants to merge 1 commit into
mainfrom
claude/hopeful-cray-jto1wh
Draft

Expect 48 files in pmm-admin summary zip (sep-provision.log)#1166
claude[bot] wants to merge 1 commit into
mainfrom
claude/hopeful-cray-jto1wh

Conversation

@claude

@claude claude Bot commented Aug 12, 2026

Copy link
Copy Markdown
Contributor

Failures fixed (investigator)

What failed

CLI / Integration / Generic was the only red job in that run, and its own
Run CLI tests step reported success — the failure surfaced one step later, when
launchable gate exited 1 on Actionable Failures | 1:

expect(received).toHaveLength(expected)
Expected length: 47
Received length: 48
  at cli/tests/generic.spec.ts:199

Root cause — a stale count in our test, not a product bug

The 48th file in the summary archive is server/sep-provision.log.

percona/pmm#5755 (PMM-15280, part of the
PMM-15205 SEP stack this FB builds) adds a sep-provision supervisord program whose stdout
goes to /srv/logs/sep-provision.log. PMM Server's logs.zip — and therefore
pmm-admin summary — now carries one more file.

It is not gated on SEP being enabled: the program is autostart = true and merely exits
early when PMM_ENABLE_SEP is unset, so supervisord creates the log either way. The CI job
starts the server with no SEP variables at all and still got 48.

That PR already updated PMM's own copy of this expectation (api-tests/server/logs_test.go
gained "sep-provision.log"); the hard-coded 47 here was the last stale copy.

Verification

Reproduced and fixed on a throwaway Linode VM following
.github/workflows/runner-integration-cli-tests.yml step for step — FB server image
perconalab/pmm-server-fb:PR-4500-38283f6 (the PR's current head build, one commit newer
than the failing run's d9f3741), the matching FB client tarball, and the same
--database pdpgsql=16 --database ps,ENCRYPTED_CLIENT_CONFIG=true setup:

  • /srv/logs/ on the FB server contains sep-provision.log with SEP disabled.

  • Before (spec at main): 1 failedExpected length: 47 / Received length: 48,
    the same assertion and the same numbers as CI. Reading the archive directly confirmed the
    delta:

    total files: 48
    sep-provision present: True
    ['server/sep-provision.log']
    
  • After (this branch, same VM): the test passes (1 passed). A full
    --grep "@generic|@unregister" run on this branch gave 43 passed, 13 skipped, 1 failed,
    the single failure being PMM-T2227 - Verify tarball upgrade reading
    process.env.PMM_CLIENT_VERSION as undefined — a variable my repro shell had not exported
    and CI does set; that test passed in the CI run this fix comes from, and this change cannot
    affect it.

The archive's other counted assertions are untouched and unaffected: the 10- and 8-file
ones are --skip-server (client-only), and the two 43-file ones sit inside permanently
test.skip(true, …) --pprof tests.

What this does and does not prove

The count is verified against a build that carries #5755. Until that PR merges, main still
produces 47 files, so this test is expected to fail on main — hence the draft status.
The gate for flipping it to ready is #5755 landing, after which the normal CLI runs confirm
it.


Generated by Claude Code

percona/pmm#5755 adds a sep-provision supervisord program whose stdout log,
/srv/logs/sep-provision.log, is collected into the server's logs.zip. The
@Generic summary test still expects 47 files, so it fails with 48 on any build
carrying that change - first seen on the PMM-15205 SEP feature build.

Signed-off-by: Claude <noreply@anthropic.com>
@travagliad travagliad added the blocked PR is Blocked to be merged label Aug 14, 2026 — with Claude
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

blocked PR is Blocked to be merged

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants