I am using Rancher 1.5.5. DNS resolution between containers is slow. It is taking 5sec to resolve the IP address of the container. It is the same delay for external alias or service link and it is fast when we use full name with rancher.internal.
search qa-116.rancher.internal qa-app1.qa-116.rancher.internal rancher.internal
What is the fix for this?
Sounds like the resolver configured in the host is unreachable and is timing out before moving on.
why it is looking in the host when it is running on managed network?
Because if you look up google.com you’d probably like an answer. Anything that doesn’t match an internal record is sent to the resolver configured in the host ('s Docker daemon).
I have service name abc which is linked to another service xyz. Why Rancher sending request to host dns to resolve abc/equivalent service link name. Is something need to do with lookup priority? If yes, how to fix it?