Help with DNS resolver
uzmargomez opened this issue · comments
Hello! I'm really new on this topic, and I'm trying to translate what I learned from the "Local playground for testing and development" example to my own 3 cluster setup (one of them with the edgeDNS resolver), these clusters are connected on my company's network. So far, this is the DNSEndpoint on an app I deployed in my 2 different clusters
Cluster 1
apiVersion: externaldns.k8s.io/v1alpha1
kind: DNSEndpoint
metadata:
annotations:
k8gb.absa.oss/dnstype: local
creationTimestamp: "2023-05-17T09:06:57Z"
generation: 2
labels:
k8gb.absa.oss/dnstype: local
name: podinfo
namespace: podinfo
ownerReferences:
- apiVersion: k8gb.absa.oss/v1beta1
blockOwnerDeletion: true
controller: true
kind: Gslb
name: podinfo
uid: abea5965-a215-43c7-8d37-d94043ee643d
resourceVersion: "80741179"
uid: 60397507-b758-4570-9aff-dab0c1f04944
spec:
endpoints:
- dnsName: localtargets-liqo.cloud.testanim.uzmar
recordTTL: 30
recordType: A
targets:
- 10.181.61.145
- 10.181.61.146
- 10.181.61.20
- 10.181.61.21
- 10.181.61.22
- dnsName: liqo.cloud.testanim.uzmar
labels:
strategy: roundRobin
recordTTL: 30
recordType: A
targets:
- 10.181.61.145
- 10.181.61.146
- 10.181.61.20
- 10.181.61.21
- 10.181.61.22
Cluster 2
kind: DNSEndpoint
metadata:
annotations:
k8gb.absa.oss/dnstype: local
creationTimestamp: "2023-05-17T09:07:02Z"
generation: 3
labels:
k8gb.absa.oss/dnstype: local
name: podinfo
namespace: podinfo
ownerReferences:
- apiVersion: k8gb.absa.oss/v1beta1
blockOwnerDeletion: true
controller: true
kind: Gslb
name: podinfo
uid: 29a4fff1-3ffd-414f-93e3-63bcc5b85b75
resourceVersion: "156700"
uid: 6a3f759b-2a49-425a-8473-63ee6700d83d
spec:
endpoints:
- dnsName: localtargets-liqo.cloud.testanim.uzmar
recordTTL: 30
recordType: A
targets:
- 10.171.61.136
- 10.171.61.148
- dnsName: liqo.cloud.testanim.uzmar
labels:
strategy: roundRobin
recordTTL: 30
recordType: A
targets:
- 10.171.61.136
- 10.171.61.148
The "localtargets-liqo.cloud.testanim.uzmar" targets correctly show the node IPs for each cluster. However, as I understand, the "liqo.cloud.testanim.uzmar" targets should show all the IPs of both of my cluster's nodes. Any idea what could be going wrong?
Also, my Edge DNS resolver running on ip 10.171.61.137 and with the 30053 port open, shows the following
dig @10.171.61.137 -p 30053 gslb-ns-eu-cloud.testanim.uzmar +short
10.181.61.146
10.181.61.21
10.181.61.22
10.181.61.145
10.181.61.20
and
dig @10.171.61.137 -p 30053 gslb-ns-us-cloud.testanim.uzmar +short
10.171.61.136
10.171.61.148
but when trying to access liqo.cloud.testanim.uzmar directly, I get the following
dig @10.171.61.137 -p 30053 liqo.cloud.testanim.uzmar
;; communications error to 10.171.61.137#30053: timed out
;; communications error to 10.171.61.137#30053: timed out
; <<>> DiG 9.18.12-0ubuntu0.22.04.1-Ubuntu <<>> @10.171.61.137 -p 30053 liqo.cloud.testanim.uzmar
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: SERVFAIL, id: 56972
;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
; COOKIE: 075cebd109008789010000006464af20a7c41b2bff774b44 (good)
;; QUESTION SECTION:
;liqo.cloud.testanim.uzmar. IN A
;; Query time: 99 msec
;; SERVER: 10.171.61.137#30053(10.171.61.137) (UDP)
;; WHEN: Wed May 17 11:40:04 BST 2023
;; MSG SIZE rcvd: 82
I wonder if there is something I may be misunderstanding. Thanks for any help or advise that you can provide!
I managed to make it work by using a Load Balancer service for the k8gb-coredns service instead of using the nginx ingress.
Hi @uzmargomez Thanks a lot for trying out the project, I am happy that it worked! I was not fast enough here but please don't hesitate to ask if you have any more questions :)
Thanks @ytsarev! No problem at all, this is a really useful and great project!