-
Task
-
Resolution: Done
-
Major
-
None
-
None
-
Quality / Stability / Reliability
-
3
-
False
-
-
False
-
-
-
3
-
Workload Mgmt Train 28 - 1, Workload Mgmt Train 28 - 2, Workload Mgmt Train 29 - 1, Workload Mgmt Train 30 - 1, Workload Mgmt Train 30 - 2
-
None
Status Update: INITIAL BM TESTING COMPLETED
Completed first comprehensive testing of RHOV spoke cluster (clc-bm-kv) on mist10 BM environment for ALC regression capacity.
📊 Test Results Summary
Environment: ACM 2.14.0 on mist10-0.qe.red-chesterfield.com with RHOV spoke cluster `clc-bm-kv`
Test Inventory: 96 total @non-ui test cases across 30 test suites
Testing Date: August 7, 2025
🎯 Actual Results
Successfully Passing: 5 test cases out of 96 total (5.2% success rate)
- Console_UI_Changes_Test_Suite: 2/2 PASSED (6m 26s)
- Git_Application_Test_Suite: 3/7 PASSED (14m 46s)
Tests Attempted but Failed: ~20 test cases
- Application_Addon_Test_Suite: 0/2 PASSED (Secret validation errors)
- Managed_Service_Account_Test_Suite: Failed (timeout issues)
- Cluster_Permission_Test_Suite: Failed (API connectivity)
- Restore_Hibernated_Cluster: Failed (function definition error)
Not Yet Tested: ~71 test cases (74% of total inventory)
🚧 Major Challenges Overcome
1. Infinite Loop Elimination
Challenge: Git/Helm tests stuck in infinite `cy:fetch ❖ STUBBED undefined undefined` loops, never completing
Root Cause: Browser-context GitHub API failures in BM environment due to CORS/CSP policies
Solution: Implemented BM-specific test variants bypassing GitHub API calls with direct input
Result: Tests now complete predictably in 6-15 minutes instead of hanging indefinitely
2. Memory Management Crisis
Challenge: Electron renderer crashes killing test runs after 15-20 minutes
Root Cause: Default Node.js memory limits insufficient for BM resource constraints
Solution: Increased Node.js memory to 8GB, optimized Cypress configuration
Result: Eliminated all memory-related test failures
3. Environment Detection
Challenge: Same test code needs to work in both cloud and BM environments
Solution: Implemented automatic BM environment detection with `CYPRESS_TEST_MODE='BM'`
Result: Smart configuration switching between cloud (GitHub API) and BM (direct input) modes
⚠️ Identified Issues Requiring Fixes
Quick Wins (4 hours total):
- Secret YAML format error: `data:` → `stringData:` conversion (30 mins)
- Function definition missing: Add `deleteNamespaceTarget` function (15 mins)
- MSA timeout adjustments: Increase BM timeout values (3 hours)
Medium Effort (20+ hours):
- API server retry logic for cluster permission tests
- BM optimization for remaining 25 untested suites
📈 Key Metrics
| Metric | BM Results |
|---|---|
| Test Completion Rate | 100% (all tests finish vs hanging) |
| Memory Stability | Stable with 8GB limit |
| Current Success Rate | 5.2% (5 of 96 tests passing) |
| Projected with Fixes | 25-35% (24-34 tests) achievable |
📋 BM Configuration
Required Setup:
export CYPRESS_TEST_MODE='BM' export NODE_OPTIONS="--max-old-space-size=8192" npx cypress run --env TEST_MODE=BM --config experimentalMemoryManagement=true
🎯 Conclusion
Current State: 5 test cases verified working on BM infrastructure
Infrastructure Capacity: RHOV spoke cluster handles testing load effectively with proper configuration
Growth Path: Clear roadmap to 25-35 working test cases with identified fixes
The BM environment can support ALC regression testing. Key breakthrough was solving test completion issues that previously prevented any meaningful testing.
📖 Technical Documentation
Complete implementation details, env setup guidelines and remediation plans:
Documentation: https://docs.google.com/document/d/10QNn5ciX9ZDIC_QbdhRFw82FUURfAsV6aoDqWmELyOE/edit?tab=t.ute0yvhp1jh4
- is related to
-
ACM-22982 Continue working on ALC BM regression
-
- New
-