0

I deployed an RKE cluster on CentOS 8 with 3 nodes.

The installation was successful but figured out that pod cannot resolve external URL because the ClusterIP of CoreDNS is not reachable.

Only pod to pod communication works.

I checked iptables rules for each Cluster IP entry and they exist.

What could be the root cause of this issue? Any hints?

Amit Kumar Gupta
  • 17,184
  • 7
  • 46
  • 64
MelDev
  • 275
  • 1
  • 6
  • 18
  • Could you share your [configuration of the RKE cluster](https://rancher.com/docs/rke/latest/en/installation/#creating-the-cluster-configuration-file)? Which [CNI plugin](https://rancher.com/docs/rancher/v2.5/en/faq/networking/cni-providers/#what-cni-providers-are-provided-by-rancher) are you using? – Mikolaj S. Dec 28 '21 at 16:20

1 Answers1

0

Change the configuration file: /etc/sysctl.conf

net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
net.bridge.bridge-nf-call-arptables = 1

Source: https://github.com/coredns/coredns/issues/1879#issuecomment-398679334

Rakesh Gupta
  • 3,507
  • 3
  • 18
  • 24
  • I have already done this steps but still have the issue, – MelDev Dec 28 '21 at 11:30
  • Also i ve set FELIX_IPTABLESBACKEND=nft and also blacklisted the ip_tables module. Firewalld is disabled too. but still have the issue, any hints please ? – MelDev Dec 28 '21 at 11:47