Trouble setting up k8gb proof of concept

[cargotrucker]

No external targets found when deploying GSLB

Hi,

I'm trying to build a K8GB proof of concept and have run into an issue where no A records are being created for my application hostname.

Environment

• 2 Kubernetes clusters (dcr1 and dcr2)
• K8GB v0.19.0
• Infoblox as edge DNS
• CoreDNS exposed through LoadBalancer services
• DNS delegation from Infoblox to the K8GB CoreDNS instances is working
• Connectivity between clusters on UDP/53 has been verified

Setup

Each cluster has a shared ingress controller configured with wildcard certificates:

• *.apps.dc1.mycorp.com
• *.apps.dc2.mycorp.com

My goal is to expose an application through K8GB using:

• podinfo.gl.k8gb.mycorp.com

The application is reachable through the ingress, but K8GB never creates any A records in CoreDNS.

Observed behavior

A DNSEndpoint resource is created, but both A records remain empty:

spec:
  endpoints:
  - dnsName: localtargets-podinfo.gl.k8gb.mycorp.com
    recordTTL: 30
    recordType: A

  - dnsName: podinfo.gl.k8gb.mycorp.com
    labels:
      strategy: failover
    recordTTL: 30
    recordType: A

The K8GB controller logs consistently show:

No external targets have been found for host host=podinfo.gl.k8gb.mycorp.com

Full relevant log excerpt:

2026-06-10T16:16:14Z INF Reconciling {"Gslb":"k8gb/podinfo"}
2026-06-10T16:16:14Z DBG Resolved strategy gslb=podinfo namespace=k8gb strategy=failover
2026-06-10T16:16:14Z INF Found Ingress Name=podinfo
2026-06-10T16:16:14Z DBG Resolved LoadBalancer and Server configuration referenced by Ingress gslb=podinfo
2026-06-10T16:16:14Z INF Adding external Gslb targets from cluster cluster=gslb-ns-dcr2-gl.k8gb.mycorp.com
2026-06-10T16:16:14Z INF Resolved glue A record for NS edgeDNSServers=10.56.76.200:53,10.56.76.192:53,10.56.140.190:53 glueARecord=[{"A":"10.56.140.190"}] nameserver=gslb-ns-dcr2-gl.k8gb.mycorp.com
2026-06-10T16:16:14Z INF No external targets have been found for host host=podinfo.gl.k8gb.mycorp.com

GSLB resource

apiVersion: k8gb.io/v1beta1
kind: Gslb
metadata:
  name: podinfo
  namespace: k8gb
spec:
  ingress: {}
  resourceRef:
    apiVersion: networking.k8s.io/v1
    kind: Ingress
    matchLabels:
      app: podinfo
  serviceHealthPolicy: Any
  strategy:
    dnsTtlSeconds: 30
    type: failover

status:
  hosts: podinfo.gl.k8gb.mycorp.com
  serviceHealth:
    podinfo.gl.k8gb.mycorp.com: Healthy

CoreDNS configuration

gl.k8gb.mycorp.com:5353 {
    import k8gbplugins
}

K8GB values

The following K8GB values are configured on the cluster:

dnsZones:
  - parentZone: "k8gb.mycorp.com"
    loadBalancedZone: "gl.k8gb.mycorp.com"
    dnsZoneNegTTL: 30
    extraPlugins: []
    extraServerBlocks: ""
    geoDataFilePath: ""
    geoDataField: ""

edgeDNSServers:
  - "10.56.76.200" # Infoblox
  - "10.56.76.192" # CoreDNS dcr1
  - "10.56.140.190" # CoreDNS dcr2

clusterGeoTag: "dcr1"
extGslbClustersGeoTags: "dcr1,dcr2"

Question

Could the shared ingress-controller setup be preventing K8GB from discovering external targets?

My expectation is that K8GB would resolve the ingress controller's external IP(s) and create A records for:

podinfo.gl.k8gb.mycorp.com

However, the controller reports that no external targets can be found.

Has anyone successfully deployed K8GB with a shared ingress controller and wildcard certificates, or is there something obvious I'm missing in my configuration?

Any pointers would be appreciated.

[ytsarev]

Hi @cargotrucker ,

Thanks for trying k8gb out! Always excited to see the new community members who are interested in the project/

    apiVersion: networking.k8s.io/v1
    kind: Ingress
    matchLabels:
      app: podinfo

can you show full reference Ingress object? Are external ips are properly set there by the ingress controller?