I have two Microservices deployed on K8S cluster (Locally on 3 VMs - 1 Master and 2 Worker Nodes):<br> 1- currency-exchange Microservice<br> 2- currency-conversion Microservice<br>
I am trying to call currency-exchange Microservice from currency-conversion by using service name : <br> http:///currency-exchange:8000.
It returns error as below: <br>
{"timestamp":"2021-02-17T08:38:25.590+0000","status":500,"error":"Internal Server Error","message":"currency-exchange executing GET http://currency-exchange:8000/currency-exchange/from/EUR/to/INR","path":"/currency-conversion/from/EUR/to/INR/quantity/10"}
I am using Kubernetes, CentOS8 using Calico CNI with set FELIX_IPTABLESBACKEND=NFT , based on this link to facilitate POD-TO-POD communications. <br> Current services available:
[root@k8s-master ~]# kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
currency-conversion NodePort 10.106.70.108 <none> 8100:32470/TCP 3h40m
currency-exchange NodePort 10.110.232.189 <none> 8000:31776/TCP 3h41mPods:<br>
[root@k8s-master ~]# kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
currency-conversion-86d9bc4698-rxdkh 1/1 Running 0 5h45m 192.168.212.125 worker-node-1 <none> <none>
currency-exchange-c79ff888b-c8sdd 1/1 Running 0 5h44m 192.168.19.160 worker-node-2 <none> <none>
currency-exchange-c79ff888b-nfqpx 1/1 Running 0 5h44m 192.168.212.65 worker-node-1 <none> <none>List of CoreDNS Pods available:<br>
[root@k8s-master ~]# kubectl get pods -o wide -n kube-system | grep coredns
coredns-74ff55c5b-9x5qm 1/1 Running 8 25d 192.168.235.218 k8s-master <none> <none>
coredns-74ff55c5b-zkkn7 1/1 Running 8 25d 192.168.235.220 k8s-master <none> <none>List all ENV variables:<br>
[root@k8s-master ~]# kubectl exec -it currency-conversion-86d9bc4698-rxdkh -- printenv
HOSTNAME=currency-conversion-86d9bc4698-rxdkh
CURRENCY_EXCHANGE_SERVICE_HOST=http://currency-exchange
KUBERNETES_SERVICE_HOST=10.96.0.1
CURRENCY_EXCHANGE_SERVICE_PORT=8000
........
nslookup kubernetes.default exec Command :<br>
[root@k8s-master ~]# kubectl exec -it currency-conversion-86d9bc4698-rxdkh -- nslookup kubernetes.default
nslookup: can't resolve '(null)': Name does not resolve
nslookup: can't resolve 'kubernetes.default': Try again
command terminated with exit code 1How do people solve such a problem? do they configure/tweak DNS to work properly as service registry ?
Thanks in advance
EDITED:
[root@k8s-master ~]# kubectl describe service currency-conversion
Name: currency-conversion
Namespace: default
Labels: app=currency-conversion
Annotations: <none>
Selector: app=currency-conversion
Type: NodePort
IP Families: <none>
IP: 10.106.70.108
IPs: 10.106.70.108
Port: <unset> 8100/TCP
TargetPort: 8100/TCP
NodePort: <unset> 32470/TCP
Endpoints: 192.168.212.125:8100
Session Affinity: None
External Traffic Policy: Cluster
Events: <none>
[root@k8s-master ~]# kubectl describe service currency-exchange
Name: currency-exchange
Namespace: default
Labels: app=currency-exchange
Annotations: <none>
Selector: app=currency-exchange
Type: NodePort
IP Families: <none>
IP: 10.110.232.189
IPs: 10.110.232.189
Port: <unset> 8000/TCP
TargetPort: 8000/TCP
NodePort: <unset> 31776/TCP
Endpoints: 192.168.19.160:8000,192.168.212.65:8000
Session Affinity: None
External Traffic Policy: Cluster
Events: <none>
I just discovered wrong behavior in all coredns PODS, a lot of timeouts:
[root@k8s-master ~]# kubectl logs coredns-74ff55c5b-zkkn7 -n kube-system
.:53
[INFO] plugin/reload: Running configuration MD5 = db32ca3650231d74073ff4cf814959a7
CoreDNS-1.7.0
linux/amd64, go1.14.4, f59c03d
[ERROR] plugin/errors: 2 6675909625369157619.3582573715596351475. HINFO: read udp 192.168.235.221:59744->192.168.100.1:53: i/o timeout
[ERROR] plugin/errors: 2 6675909625369157619.3582573715596351475. HINFO: read udp 192.168.235.221:53400->192.168.100.1:53: i/o timeout
[ERROR] plugin/errors: 2 6675909625369157619.3582573715596351475. HINFO: read udp 192.168.235.221:58465->192.168.100.1:53: i/o timeout
[ERROR] plugin/errors: 2 6675909625369157619.3582573715596351475. HINFO: read udp 192.168.235.221:58197->192.168.100.1:53: i/o timeout
[ERROR] plugin/errors: 2 6675909625369157619.3582573715596351475. HINFO: read udp 192.168.235.221:57794->192.168.100.1:53: i/o timeout
[ERROR] plugin/errors: 2 6675909625369157619.3582573715596351475. HINFO: read udp 192.168.235.221:43345->192.168.100.1:53: i/o timeout
[ERROR] plugin/errors: 2 6675909625369157619.3582573715596351475. HINFO: read udp 192.168.235.221:57361->192.168.100.1:53: i/o timeout
[ERROR] plugin/errors: 2 6675909625369157619.3582573715596351475. HINFO: read udp 192.168.235.221:51716->192.168.100.1:53: i/o timeouthow can i start trace the problem?
Extra Details:
[root@k8s-master ~]# kubectl exec -i -t currency-conversion-86d9bc4698-rxdkh -- sh
/ # wget http://currency-exchange:8000/currency-exchange/from/EUR/to/INR
wget: bad address 'currency-exchange:8000'It looks to me that you have incorrectly setup your CNI overlay network. I checked your previous question to verify node's ip address and to me it looks that your pod network overlap with your host network:
The Kubernetes pod-network-cidr is the IP prefix for all pods in the Kubernetes cluster. This range must not clash with other networks in your VPC
The Kubernetes pod network documentation describes this as well:
Take care that your Pod network must not overlap with any of the host networks: you are likely to see problems if there is any overlap. (If you find a collision between your network plugin's preferred Pod network and some of your host networks, you should think of a suitable CIDR block to use instead, then use that during
kubeadm initwith--pod-network-cidrand as a replacement in your network plugin's YAML).
This is also mentioned in calico instructions when creating a cluster:
Note: If 192.168.0.0/16 is already in use within your network you must select a different pod network CIDR, replacing 192.168.0.0/16 in the above command.
PS. You can always wget curl from here.