- Newest
- Most votes
- Most comments
Hello,
Thank you for providing all the requested information. I really appreciate it!
So you have 5 pods running at the moment and all of them are running in Fargate. As per the below information, you have 2 core-dns pods and 3 pods in flink-session namespace all running in Fargate.
Since you are running **all the pods **in Fargate, you do not need to run kube-proxy separately as they are automatically deployed upon creation of the fargate nodes and they’re fully managed by AWS. Please note that Fargate adds 256 MB to each Pod's memory reservation for the required Kubernetes components in all Fargate nodes(kubelet, kube-proxy, and containerd).
If you are planning to only use Fargate nodes, you can safely remove kube-proxy add-on.
NAMESPACE NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
flink-session flink-jobmanager-69569d5ff5-fl7z5 1/1 Running 0 3h7m 172.31.52.xx fargate-ip-172-31-52-49.us-east-2.compute.internal <none> <none>
flink-session flink-taskmanager-755b89f9dc-9pcqh 1/1 Running 0 3h7m 172.31.62.xx fargate-ip-172-31-62-13.us-east-2.compute.internal <none> <none>
flink-session flink-taskmanager-755b89f9dc-zphxm 1/1 Running 0 3h7m 172.31.62.xx fargate-ip-172-31-62-98.us-east-2.compute.internal <none> <none>
kube-system coredns-6748b94b5c-b62x7 1/1 Running 0 5d4h 172.31.62.xxx fargate-ip-172-31-62-158.us-east-2.compute.internal <none> <none>
kube-system coredns-6748b94b5c-cc2n9 1/1 Running 0 5d4h 172.31.61.xx fargate-ip-172-31-61-43.us-east-2.compute.internal <none> <none> kube-system kube-proxy-ds94h 0/1 Pending 0 128m <none> <none> <none> <none>
kube-system kube-proxy-f76f9 0/1 Pending 0 128m <none> <none> <none> <none>
kube-system kube-proxy-pksz6 0/1 Pending 0 128m <none> <none> <none> <none>
kube-system kube-proxy-vp2bv 0/1 Pending 0 128m <none> <none> <none> <none>
kube-system kube-proxy-zxf65 0/1 Pending 0 128m <none> <none> <none>
Please let me know in case of any further queries.
Thanks, Manish
answered 3 years ago
Thanks。The service exposed with nodeIp:port.
Relevant content
asked 2 years ago
asked 10 months ago
Hello, Could you please some more information about you cluster? How many nodes you have in your cluster? Logs of kube-proxy pod and what all pods are in running in kube-system namespace? pod, service and ingress manifest, if possible.
hello, cluster is fargate type EKS. There are 5 nodes and the status are ready. So I deploy a flink cluster on this cluster,and running well. There is no log about kube-proxy pod, the pod is pending ,so there no log generate. And I describe a kube-proxy pod show the warning : Warning FailedScheduling 38s default-scheduler 0/5 nodes are available: 1 Too many pods. preemption: 0/5 nodes are available: 5 No preemption victims found for incoming pod.
and the pod are pending always. and describe a node: CPU Requests: 100m (5%) Limit: 1 (50%). MEM: 1Gi (27%) LIMIT: 2Gi (54%) kubelet has sufficient memory。
namespace： kube-system, default, stream, flink-session Service：service/flink-jobmanager。service/flink-jobmanager-rest and service/kube-dns. I have not install ingress.
fargate pattern: Daemonsets aren't supported on Fargate. If your application requires a daemon, reconfigure that daemon to run as a sidecar container in your Pods. (doc from AWS https://docs.aws.amazon.com/eks/latest/userguide/fargate.html)
NAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE daemonset.apps/aws-node 0 0 0 0 0 <none> 15d daemonset.apps/kube-proxy 5 5 0 5 0 eks.amazonaws.com/compute-type=fargate,kubernetes.io/os=linux 12