Summary

This PR improves the reliability and debuggability of our e2e test suite by:

1. Standardizing bash error handling across all 47 e2e chainsaw tests by changing set -e to set -euo pipefail
2. Adding CCM log collection on test failure to help debug issues

Changes

Bash Error Handling Standardization

• Replaced set -e with set -euo pipefail in all script blocks
• set -euo pipefail provides:
  • -e: Exit on any command failure
  • -u: Treat unset variables as errors
  • -o pipefail: Exit if any command in a pipeline fails

CCM Log Collection on Failure

• Added catch blocks to all 47 e2e tests
• On test failure, automatically fetches the last 100 lines of CCM logs filtered by test namespace
• Uses kubectl logs -n kube-system daemonsets/ccm-linode | grep "<test-namespace>" | tail -100

Impact

• Better error detection: Tests will fail more reliably on subtle issues like unset variables or pipeline failures
• Improved debugging: Test failures now automatically include relevant CCM logs for faster troubleshooting
• Consistent behavior: All e2e tests now use the same robust bash error handling

Files Changed

• 47 e2e chainsaw test files across e2e/test/, e2e/bgp-test/, and e2e/subnet-test/

Testing

• Changes are purely infrastructure improvements to the test framework
• No functional changes to the CCM itself
• All existing e2e tests should continue to pass with improved error detection