README.md

node-local-dns

NodeLocal DNS Cache helm chart

Installing the Chart

This chart deploys NodeLocal DNSCache Daemon set according to https://kubernetes.io/docs/tasks/administer-cluster/nodelocaldns/.

It is designed to work both with iptables and IPVS setup.

Latest available node-local-dns image can be found at node-local-dns google container repository

Values

Key	Type	Default	Description
affinity	object	{}
config.localDnsIp	string	"169.254.20.11"
config.zones.".:53".plugins.cache.denial	object	{}
config.zones.".:53".plugins.cache.parameters	int	30
config.zones.".:53".plugins.cache.prefetch	object	{}
config.zones.".:53".plugins.cache.serve_stale	bool	false
config.zones.".:53".plugins.cache.success	object	{}
config.zones.".:53".plugins.debug	bool	false
config.zones.".:53".plugins.errors	bool	true
config.zones.".:53".plugins.forward.except	string	""
config.zones.".:53".plugins.forward.expire	string	""
config.zones.".:53".plugins.forward.force_tcp	bool	false
config.zones.".:53".plugins.forward.health_check	string	""
config.zones.".:53".plugins.forward.max_fails	string	""
config.zones.".:53".plugins.forward.parameters	string	"__PILLAR__UPSTREAM__SERVERS__"
config.zones.".:53".plugins.forward.policy	string	""
config.zones.".:53".plugins.forward.prefer_udp	bool	false
config.zones.".:53".plugins.health.port	int	8080
config.zones.".:53".plugins.log.classes	string	"all"
config.zones.".:53".plugins.log.format	string	"combined"
config.zones.".:53".plugins.prometheus	bool	true
config.zones.".:53".plugins.reload	bool	true
config.zones.".:53".plugins.template	object	{}
config.zones."in-addr.arpa:53".plugins.cache.parameters	int	30
config.zones."in-addr.arpa:53".plugins.debug	bool	false
config.zones."in-addr.arpa:53".plugins.errors	bool	true
config.zones."in-addr.arpa:53".plugins.forward.force_tcp	bool	false
config.zones."in-addr.arpa:53".plugins.forward.parameters	string	"__PILLAR__UPSTREAM__SERVERS__"
config.zones."in-addr.arpa:53".plugins.health.port	int	8080
config.zones."in-addr.arpa:53".plugins.log.classes	string	"all"
config.zones."in-addr.arpa:53".plugins.log.format	string	"combined"
config.zones."in-addr.arpa:53".plugins.prometheus	bool	true
config.zones."in-addr.arpa:53".plugins.reload	bool	true
config.zones."ip6.arpa:53".plugins.cache.parameters	int	30
config.zones."ip6.arpa:53".plugins.debug	bool	false
config.zones."ip6.arpa:53".plugins.errors	bool	true
config.zones."ip6.arpa:53".plugins.forward.force_tcp	bool	false
config.zones."ip6.arpa:53".plugins.forward.parameters	string	"__PILLAR__UPSTREAM__SERVERS__"
config.zones."ip6.arpa:53".plugins.health.port	int	8080
config.zones."ip6.arpa:53".plugins.log.classes	string	"all"
config.zones."ip6.arpa:53".plugins.log.format	string	"combined"
config.zones."ip6.arpa:53".plugins.prometheus	bool	true
config.zones."ip6.arpa:53".plugins.reload	bool	true
image.args.healthPort	int	8080
image.args.interfaceName	string	"nodelocaldns"
image.args.quiet	bool	false
image.args.setupInterface	bool	true
image.args.setupIptables	bool	false
image.args.skipTeardown	bool	true
image.args.syncInterval	string	"1ns"
image.args.upstreamSvc	string	"kube-dns"
image.pullPolicy	string	"IfNotPresent"
image.repository	string	"registry.k8s.io/dns/k8s-dns-node-cache"
image.tag	string	"1.23.0"
imagePullSecrets	list	[]
metrics.port	int	9253
metrics.prometheusScrape	string	"true"
nodeSelector	object	{}
podAnnotations	object	{}
podSecurityContext	object	{}
podmonitor.enabled	bool	false
podmonitor.metricRelabelings	list	[]
priorityClassName	string	"system-node-critical"
readinessProbe	string	nil
resources.requests.cpu	string	"30m"
resources.requests.memory	string	"50Mi"
securityContext.privileged	bool	true
serviceAccount.annotations	object	{}
serviceAccount.create	bool	true
serviceAccount.name	string	""
tolerations[0].key	string	"CriticalAddonsOnly"
tolerations[0].operator	string	"Exists"
tolerations[1].effect	string	"NoExecute"
tolerations[1].operator	string	"Exists"
tolerations[2].effect	string	"NoSchedule"
tolerations[2].operator	string	"Exists"
updateStrategy.rollingUpdate.maxUnavailable	string	"10%"
useHostNetwork	bool	true

Additional Information

Cilium

For clusters running cilium, there is a CRD, local-redirect-policy, which needs be extra enabled via --set localRedirectPolicy=true. It enables pod traffic destined to an IP address and port/protocol tuple or Kubernetes service to be redirected locally to backend pod(s) within a node, using eBPF. The namespace of backend pod(s) need to match with that of the policy.

For using this feature, values should provides the following extra configuration,

For getting the CLUSTER_DNS_IP,

kubectl get svc kube-dns -n kube-system -o jsonpath={.spec.clusterIP}

config:
  localDnsIp: CLUSTER_DNS_IP
  cilium:
    clusterDNSService: kube-dns
    clusterDNSNamespace: kube-system
    udp:
      enabled: true
      portName: dns
    tcp:
      enabled: true
      portName: dns-tcp

RKE2

As this feature heavily depends on the Cluster DNS implementation, for a Rancher Kubernetes Engine 2 cluster, clusterDNSService should be rke2-coredns-rke2-coredns, and port names, udp-53 and tcp-53 respectively.

---

Autogenerated from chart metadata using helm-docs v1.14.2