HTTP routes error if a node goes offline #7842
Description
Steps to reproduce
- Have a multi-node NethServer 8 cluster.
- Take one node offline (shutdown or disconnect it).
- Access the cluster-manager on any of the remaining nodes and go to Settings > HTTP routes.
- Try to list routes for "Any node" or for the offline node.
Expected behavior
The HTTP routes menu should display all available routes, skipping or properly informing about unreachable nodes. The presence of an offline node should not cause an error for the entire menu.
Actual behavior
If at least one node is offline, the HTTP routes menu in cluster-manager on the cluster leader displays an error message:
This is likely because the APIs of the traefik module on the offline node are unreachable. Example command results:
~]# api-cli run module/traefik2/list-routes
["cluster-admin", "openldap2-amld"]
~]# api-cli run module/traefik1/list-routes
TaskSubmissionCheckFailed: Client "module/traefik1" was not found
As a workaround, remove the offline node, as described in the manual.
Components
- NethServer 8 cluster-manager
- Multi-node cluster with at least one node offline
- Version and RPM details (see Software Center or run
rpm -qa | grep -F .ns8. | sort)
See also
- This issue appears during (but is not limited to) switch-leader operations.
- Context: https://mattermost.nethesis.it/nethesis/pl/wmt6edmuypy78k8qz9jpsarahc
Thanks to @nrauso