Skip to content

debug-nexus-stability.md

Latest commit

 

History

History
37 lines (29 loc) · 1.75 KB

debug-nexus-stability.md

File metadata and controls

37 lines (29 loc) · 1.75 KB

[OPEN] Nexus stability debugging session

Session ID: nexus-stability Date: 2026-06-10 Project Root: /Users/stephenwang/Documents/trae_projects/DevOps

Symptom:

  • Full chain build reaches GitLab and SonarQube successfully.
  • Nexus upload succeeds in isolation.
  • After GitLab and SonarQube are both running, Nexus on port 8082 becomes unreachable during the formal Jenkins build.

Expected:

  • Nexus stays reachable long enough for arithmetic-demo-build to upload the artifact.

Hypotheses:

  1. The temporary Nexus container is OOM-killed after GitLab and SonarQube reclaim memory.
  2. Nexus remains running, but host port 8082 is no longer reachable because of container/network state.
  3. Service startup order and overlap cause Nexus to fail before it reaches a usable state.
  4. Jenkins probing is too strict or checks the wrong readiness signal for Nexus availability.

Evidence Plan:

  • Inspect Docker container state, restart counts, and OOM flags for Nexus, GitLab, SonarQube, and Jenkins.
  • Check live reachability from host and, if possible, from the Jenkins context.
  • Review recent Nexus logs around the latest failure window.
  • Reproduce one more failure with minimal overlap changes only after baseline evidence is collected.

Status:

  • Session opened.
  • No business logic changed yet.

Evidence Collected:

  • devops-nexus-fresh-test previously exited with exitCode=137 and oomKilled=true while GitLab + SonarQube were running.
  • After tuning GitLab/SonarQube and adjusting readiness checks, Nexus stayed reachable long enough for a full build to complete.

Current Verification: