v0.0.15

This patch release has critical fix that affects distros with strict reverse path filtering. When nodes span across multiple subnets, Kube-router uses ip-in-ip tunneling. Fix in this release ensures traffic originates and leaves on same tunnel interface.

DaemonSet Upgrade Steps

image: cloudnativelabs/kube-router[:latest]

imagePullPolicy: Always

These are the default options in our manifests. In this scenario you just need to delete the pods so they are recreated. This will not interrupt cluster services/traffic.

kubectl -n kube-system delete pods -l k8s-app=kube-router

image: cloudnativelabs/kube-router:vX.X.X

If your image option is pinned to a specific version, you can upgrade using kubectl set image.

kubectl -n kube-system set image ds/kube-router kube-router=cloudnativelabs/kube-router:v0.0.15
kubectl -n kube-system delete pods -l k8s-app=kube-router

Changelog

c9c6107 Merge pull request #143 from cloudnativelabs/pbr_tunnels
7b9eedc Setup policy-based routing so that traffic arriving on a tunnel interface leaves on same tunnel interface irrespective of rp_filter value
22330ab Merge pull request #146 from cloudnativelabs/goreportcard
ed935a4 fix gofmt, go_vet, misspells
b171f22 Update README.md

---

Automated with GoReleaser
Built with go version go1.8.3 linux/amd64