Trow unreachable over IPv6
Describe the bug
When running a cluster with DualStack/IPv6 trow is not reachable as it only listens on 0.0.0.0 (all IPv4 interfaces).
This could be mitigated by letting Trow/Rocket listen by default on :: (actually all interfaces IPv4 and IPv6).
Another workaround would be to allow the helm chart to set the --host flag.
To Reproduce
- Setup a cluster with IPv6 support
- Setup an ingress controller with IPv6 support
- Setup trow + ingress (or IPv6 service)
Expected behavior
Trow is reachable at the configured domain through the ingress/directly via IPv6
Trow Info
- Install method: Helm
- Helm Chart info:
apiVersion: v2
appVersion: 0.3.5
description: Helm chart for Trow registry
name: trow
type: application
version: 0.3.5
Kubernetes
- Kubernetes distro/host: kubeadm on Arch
- Kubernetes version: v1.25.5
- Container Runtime: ContainerD
Note that Trow uses Rocket, which does not support listening to multiple addresses in a single instance. The workaround described here might work tho: https://github.com/SergioBenitez/Rocket/issues/1581. Otherwise the solution is to have multiple instances of rocket running (like I did here).
Thanks @awoimbee !