Skip to content

Pre-built image network failure handling #8021

@DavidePrincipi

Description

@DavidePrincipi

The NS8 pre-built image fails irreversibly if the network is down or misconfigured. Traefik install fails and cluster-admin becomes unreachable. Even if Traefik install succeedes, a network issue may suddenly break the create-cluster action, with no recover path.

Proposed solution

  • Install Traefik in the pre-built image, finish its configuration (i.e. secrets and certs generation) at first boot, when network may be down.
  • Handle validation error of Create Cluster

Alternative solutions

Install core modules from a local Podman image cache. A local registry service should provide the required images with zero network connections.

Additional context

If "Create cluster" fails for a network/DNS failure, it cannot be executed again (VPN address already set).

Image

See also

Metadata

Metadata

Labels

No labels
No labels
No fields configured for Feature.

Projects

Status

In Progress

Relationships

None yet

Development

No branches or pull requests

Issue actions