You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
feat: add GitHub Actions network tuning to fix CI connectivity issues
- Add smorimoto/tune-github-hosted-runner-network action to E2E and LXD test workflows
- Fix intermittent network connectivity problems in GitHub Actions runners
- Address Docker GPG key download failures and apt repository timeouts
- Update documentation with comprehensive CI network troubleshooting guide
- Reference specific GitHub issues #1187 and #2890 for transparency
- All linters pass and documentation follows project standards
This should resolve the persistent 'Network is unreachable' errors in CI runs.
Copy file name to clipboardExpand all lines: docs/e2e-testing.md
+25Lines changed: 25 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -127,6 +127,31 @@ tofu destroy -auto-approve
127
127
-**SSH connectivity failures**: Usually means cloud-init is still running or SSH configuration failed
128
128
-**Ansible connection errors**: Check if the container IP is accessible and SSH key permissions are correct
129
129
-**OpenTofu errors**: Ensure LXD is properly configured and you have sufficient privileges
130
+
-**Network connectivity issues in CI**: See [CI Network Issues](#ci-network-issues) section below
131
+
132
+
### CI Network Issues
133
+
134
+
GitHub Actions runners sometimes experience intermittent network connectivity problems that can cause:
135
+
136
+
- Docker GPG key downloads to fail (`Network is unreachable` errors)
137
+
- Package repository access timeouts
138
+
- Generally flaky network behavior
139
+
140
+
**Root Cause**: This is a known issue with GitHub-hosted runners running in Azure:
141
+
142
+
-[GitHub Issue #1187](https://github.com/actions/runner-images/issues/1187) - Original networking issue
143
+
-[GitHub Issue #2890](https://github.com/actions/runner-images/issues/2890) - Specific apt repository timeout issues
144
+
145
+
**Solution**: We use the [`smorimoto/tune-github-hosted-runner-network`](https://github.com/marketplace/actions/tune-github-hosted-runner-network) action to disable TCP/UDP offload and fix these networking issues.
146
+
147
+
**Implementation**: The action is automatically added to CI workflows and runs before any network-dependent operations.
148
+
149
+
**Playbook Adaptations**: Our Ansible playbooks also include:
150
+
151
+
- CI environment detection
152
+
- Adaptive retry strategies with longer timeouts in CI
0 commit comments