Skip to content

Bound the first-member wait in OpenBao configuration #6360

Open

Description

@devantler

🤖 Generated by the Agentic Engineer

Description

A configuration prerequisite can keep the production delivery lane occupied after its deployment health check has already expired. The OpenBao initialization script waits for the designated first member in an until loop without its own attempt or elapsed-time bound. The Job has a one-hour deadline, while the infrastructure health check expires after 20 minutes.

Steps to Reproduce

  1. Exercise the initialization script in a disposable environment with the designated first member unavailable.
  2. Keep the status command returning an unavailable observation.
  3. Observe that the initial wait continues past the enclosing deployment health-check window; the script has no earlier terminal result.

Expected Behavior

The prerequisite fails closed within an explicit bound shorter than the enclosing health-check deadline, with a useful non-secret diagnostic. A later deployment can re-evaluate it after the underlying availability problem is resolved.

Actual Behavior and Evidence

The reviewed initialization script has an unbounded initial status loop. On 15 November, read-only observations confirmed that a configuration attempt was still in that wait after the infrastructure health check expired. Detailed operational evidence stays private.

Acceptance Criteria

  • Give the first-member availability wait an explicit finite bound below the enclosing health-check deadline, accounting for individual command timeouts.
  • A regression exercises the actual extracted shell script with an unavailable first member and proves a nonzero terminal result within the bound.
  • Cover successful availability before expiry and unavailable or malformed observations; expiry must perform no initialization and print no credentials.
  • Preserve the existing initialization, stored-key and snapshot recovery guards. A missing member never authorizes re-initialization or counts as quorum proof.
  • The fix passes native validation and exact-head review, and its eventual production result is observed separately from source merge.

Additional Context

Part of #4979. Related recovery and upgrade-proof work: #4756. This is a bounded prerequisite-failure fix; the active storage recovery retains its existing owner and scheduling fence.

Audience: platform operator and every change waiting for deployment. Priority rationale: a repeatable prerequisite wait consumes the delivery lane even after failure is already known. Estimated effort: Small, covering the script, behavioral regression and validation. New intake remains Backlog until refined; no competing production recovery is started.

Activity

  1. added theissue type on Nov 15, 2026
  2. added 2 commits that reference this issue on Nov 15, 2026
  3. added a commit that references this issue on Nov 15, 2026
  4. moved this to 📥 Backlog in 🌊 Project Boardon Nov 15, 2026
  5. added 2 commits that reference this issue on Nov 15, 2026
  6. moved this from 📥 Backlog to 🏃🏻‍♂️ In Progress in 🌊 Project Boardon Nov 15, 2026
  7. added 2 commits that reference this issue on Nov 15, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    Fields

    Priority

    None yet

    Effort

    None yet

    Projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions