dns issues in OKD4.9/4.10(Failed unit NetworkManager-wait-online.service) #1149
-
Hi, Provisioning/bootstrapping OKD4.9 or 4.10 from scratch on Openstack(UPI) goes pretty smooth. Boot-times of the fedoracoreOS-nodes can however take up til 2 minutes or longer(which is way too long),and often we experience also slow or not working DNS resolving. I assume both issues are connected. (a plain fedorecoreOS(not OKD) boots under 30 seconds...) On the hosts(master/worker nodes) themselves we see that default /etc/resolv.conf after provisioning of the cluster:
We assume that removing 10.10.196.15 can possibly fix our issue. The question now is if this is wise? If so, how to proceed? How can this be changed in an elegant way? default /etc/coredns/Corefile after provisioning of the cluster:
Or can |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
It should not be a local address, see #978 |
Beta Was this translation helpful? Give feedback.
It should not be a local address, see #978