What happened:
I created 2 deployments with services in the same namespace - fabrikamfiber.web and db. fabrikamfiber.web can't resolve db without using the FQDN
How to reproduce it (as minimally and precisely as possible):
Create 2 pods in the same namespace
Expose a service on one of them
Attempt to connect to it by name from the other pod
Orchestrator and version (e.g. Kubernetes, DC/OS, Swarm) Master: Kubernetes v1.7.9
Nodes Kernel Version: 10.0 16299 (16299.15.amd64fre.rs3_release.170928-1534) OS Image: Operating System: windows Architecture: amd64 Container Runtime Version: docker://Unknown Kubelet Version: v1.7.9-29+3911e53660eccc Kube-Proxy Version: v1.7.9-29+3911e53660eccc
What happened: I created 2 deployments with services in the same namespace - fabrikamfiber.web and db. fabrikamfiber.web can't resolve
db
without using the FQDNHow to reproduce it (as minimally and precisely as possible): Create 2 pods in the same namespace Expose a service on one of them Attempt to connect to it by name from the other pod
Anything else we need to know: Root cause tracked at https://github.com/Microsoft/SDN/issues/150 . ACS-Engine will need to pick up new version of wincni.exe to fix this