This section explains how to troubleshoot OpenPERouter deployments.

Inspect Tool #

The inspect tool makes debugging OpenPERouter deployments easier by collecting related objects and logs into a single directory, for inspection or attached to a bug report.

It can be found at OpenPERouter repository tools/inspect

Output #

The tool wire all collected info to output directory (default is openperouter-inspect/), including:

• node_info/ - per-node network and routing infrastructure information (IP link state, VRFs, type 2 and 5 routes, etc.)
• <openperouter namespace>/ - namespace objects and workload logs (pod logs, events, CRs, roles, etc.)
• <namespace name>/ - per-namespace directory with config resources (Underlay, L3VNI, L2VNI, etc.)

Please refer to the tool README for more details.

Systemd mode #

When OpenPERouter runs in systemd mode, the router state can’t be collected through the cluster API because the router container isn’t managed by Kubernetes.

The inspect_host tool can be used for inspecting such nodes.

It can be found at OpenPERouter repository tools/inspect

Output #

The tool store all collected info to output directory (default is /openperouter-inspect-host), including:

• router_info_podman.log - router infrastructure information collected from podman (for Podman Quadlet containers)
• router_info_crictl.log - router infrastructure information collected using crictl (e.g.: from CRI-O, containerd)
• root_netns_info.log - root network namespace information
• config_files.log - static config resources collection log
• configs/ - collected static config resources, YAML form

Please see the tool README for more details.