• Problem: I'm seeing a well-known problem of crash-looping core-dns, and i'm forwarding to

/etc/resolv.conf

which indeeds contains

nameserver 127.0.0.1

. If i manually removed the "loop" from config map, the core-dns pod would start fine, but after config map changes, it would crash soon after.. • Why can't i remove the localhost loopback? The localhost is actually intended, since I want to forward to localhost where my upstream DNS process is running on the machine node. • Question: Is there a way to distinguish between pod's localhost from node's localhost? • What i tried: adding

127.0.0.2 localhost

in it's /etc/hosts equivalent, and then i'd forward resolution to

/etc/resolv.conf

with

nameserver 127.0.0.2

seems like there's no DNS resolution happening