본문 바로가기
Container/Kubernetes

[K8S] DaemonSet

융또융 2024. 2. 26. 15:52
DaemonSet
전체 노드에서 특정 Pod가 한 개씩 실행되도록 보장
각 노드에 동일한 워크로드 (ex: 로그 수입기, 모니터링 에이전트)와 같은 프로그램 실행 시 적용

 

DaemonSet의 주요 특징

  1. 모든 노드에서 실행 보장:
    • DaemonSet은 클러스터의 모든 노드에 동일한 Pod를 하나씩 실행 한다.
    • 새 노드가 클러스터에 추가되면 DaemonSet은 자동으로 해당 노드에 Pod를 생성시킨다.
  2. 특정 노드에만 실행 가능:
    • 라벨과 Node Selector를 사용해 특정 노드에서만 실행하도록 제한 가능.
  3. 노드에서 Pod 제거:
    • 노드가 삭제되거나 DaemonSet이 수정되면 Pod도 자동으로 제거.
  4. 자동 관리:
    • 노드 추가, 삭제, 스케줄링 등 모든 작업을 DaemonSet이 자동으로 처리.

DaemonSet의 사용 사례

  1. 로그 수집:
    • 각 노드에서 로그를 수집하여 중앙 집중화된 로깅 시스템으로 전달.
      • 예: Fluentd, Logstash.
  2. 모니터링 에이전트:
    • 노드와 컨테이너 상태를 모니터링하는 에이전트 실행.
      • 예: Prometheus Node Exporter, Datadog Agent.
  3. 네트워킹:
    • 네트워크 구성 또는 라우팅 작업을 처리하는 Pod 실행.
      • 예: CNI 플러그인, DNS Resolver.
  4. 노드별 작업:
    • 모든 노드에서 반드시 실행해야 하는 특정 작업 수행.
      • 예: 보안 검사 에이전트.

 

https://www.youtube.com/watch?v=wJeb561CMOg

 

 

deploy된 모든 DaemonSet 확인

[master ~]$ kubectl get ds -A
NAMESPACE      NAME              DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR            AGE
kube-flannel   kube-flannel-ds   1         1         1       1            1           <none>                   3m33s
kube-system    kube-proxy        1         1         1       1            1           kubernetes.io/os=linux   3m34s

[master ~]$ kubectl describe daemonsets.apps kube-flannel-ds -n kube-flannel
Name:           kube-flannel-ds
Selector:       app=flannel,k8s-app=flannel
Node-Selector:  <none>
Labels:         app=flannel
                k8s-app=flannel
                tier=node
Annotations:    deprecated.daemonset.template.generation: 1
Desired Number of Nodes Scheduled: 1
Current Number of Nodes Scheduled: 1
Number of Nodes Scheduled with Up-to-date Pods: 1
Number of Nodes Scheduled with Available Pods: 1
Number of Nodes Misscheduled: 0
Pods Status:  1 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:           app=flannel
                    k8s-app=flannel
                    tier=node
  Service Account:  flannel
  Init Containers:
   install-cni-plugin:
    Image:      docker.io/flannel/flannel-cni-plugin:v1.2.0
    Port:       <none>
    Host Port:  <none>
    Command:
      cp
    Args:
      -f
      /flannel
      /opt/cni/bin/flannel
    Environment:  <none>
    Mounts:
      /opt/cni/bin from cni-plugin (rw)
   install-cni:
    Image:      docker.io/flannel/flannel:v0.23.0
[master ~]$ kubectl get nodes
NAME    STATUS   ROLES           AGE     VERSION
master   Ready   control-plane   3h27m   v1.27.2
node2   Ready    <none>          132m    v1.27.2
node3   Ready    <none>          42m     v1.27.2
node4   Ready    <none>          42m     v1.27.2

# 워커 노드 삭제
[master ~]$ kubectl delete nodes node4
node "node4" deleted
[master ~]$ kubectl get nodes
NAME    STATUS   ROLES           AGE     VERSION
master   Ready    control-plane   3h35m   v1.27.2
node2   Ready    <none>          140m    v1.27.2
node3   Ready    <none>          50m     v1.27.2

[master ~]$ cat > daemonset-exam.yaml
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: daemonset-nginx
spec:
  selector:
    matchLabels:
      app: webui
  template:
    metadata:
      name: nginx-pod
      labels:
        app: webui
    spec:
      containers:
      - name: nginx-container
        image: nginx:1.14

[master ~]$ kubectl create -f daemonset-exam.yaml 
daemonset.apps/daemonset-nginx created
[master ~]$ kubectl get pods -o wide
NAME                    READY   STATUS    RESTARTS   AGE   IP         NODE    NOMINATED NODE   READINESS GATES
daemonset-nginx-ppgll   1/1     Running   0          12s   10.5.2.5   node3   <none>           <none>
daemonset-nginx-wntfs   1/1     Running   0          12s   10.5.1.5   node2   <none>           <none>

 

#  워커노드 다시 추가하기

# token 생성
# kubeadm token create --tth [시간]
[master ~]$ kubeadm token list
TOKEN                     TTL         EXPIRES                USAGES                   DESCRIPTION                                                EXTRA GROUPS
c42htv.p6ltpxh8m2tjwghu   20h         2024-02-27T02:56:44Z   authentication,signing   The default bootstrap token generated by 'kubeadm init'.   system:bootstrappers:kubeadm:default-node-token
[master ~]$ kubeadm token create --ttl 1h
vkytah.x0y8bgq16ihf8v5z

[master ~]$ kubeadm token list
TOKEN                     TTL         EXPIRES                USAGES                   DESCRIPTION                                                EXTRA GROUPS
c42htv.p6ltpxh8m2tjwghu   20h         2024-02-27T02:56:44Z   authentication,signing   The default bootstrap token generated by 'kubeadm init'.   system:bootstrappers:kubeadm:default-node-token
vkytah.x0y8bgq16ihf8v5z   59m         2024-02-26T07:36:18Z   authentication,signing   <none>                                                     system:bootstrappers:kubeadm:default-node-token


# node4에서 kubeadm reset 실행
[node4 ~]$ kubeadm reset
W0226 06:37:47.439686   26715 preflight.go:56] [reset] WARNING: Changes made to this host by 'kubeadm init' or 'kubeadm join' will be reverted.
[reset] Are you sure you want to proceed? [y/N]: y


...
Run 'kubectl get nodes' on the control-plane to see this node join the cluster.

[node4 ~]$ kubeadm join 192.168.0.18:6443 --token vkytah.x0y8bgq16ihf8v5z   --discovery-token-ca-cert-hash sha256:7099406672b8b15a79d1286859b7a1c29a3797afa8d1a81a8d6e0dbd182a0396
W0226 06:40:54.273866     582 initconfiguration.go:120] Usage of CRI endpoints without URL scheme is deprecated and can cause kubelet errors in the future. Automatically prepending scheme "unix" to the "criSocket" with value "/run/docker/containerd/containerd.sock". Please update your configuration!

# 다시 node4 추가된 것 확인
[master ~]$ kubectl get nodes
NAME    STATUS   ROLES           AGE     VERSION
master   Ready    control-plane   3h44m   v1.27.2
node2   Ready    <none>          150m    v1.27.2
node3   Ready    <none>          60m     v1.27.2
node4   Ready    <none>          41s     v1.27.2

# 노드당 1개의 Pod 보장! ## 중요
[master ~]$ kubectl get pods -o wide
NAME                    READY   STATUS    RESTARTS   AGE     IP         NODE    NOMINATED NODE   READINESS GATES
daemonset-nginx-7frwz   1/1     Running   0          41s     10.5.4.3   node4   <none>           <none>
daemonset-nginx-ppgll   1/1     Running   0          8m28s   10.5.2.5   node3   <none>           <none>
daemonset-nginx-wntfs   1/1     Running   0          8m28s   10.5.1.5   node2   <none>           <none>

[master ~]$ kubectl delete pod daemonset-nginx-7frwz
pod "daemonset-nginx-7frwz" deleted
[master ~]$ kubectl get pods -o wide
NAME                    READY   STATUS    RESTARTS   AGE   IP         NODE    NOMINATED NODE   READINESS GATES
daemonset-nginx-2fnhc   1/1     Running   0          3s    10.5.4.4   node4   <none>           <none>
daemonset-nginx-ppgll   1/1     Running   0          10m   10.5.2.5   node3   <none>           <none>
daemonset-nginx-wntfs   1/1     Running   0          10m   10.5.1.5   node2   <none>           <none>
# 즉, 로그/ 모니터링 에이전트에 적합한 컨트롤러다.

[master ~]$ kubectl get daemonsets.apps 
NAME              DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
daemonset-nginx   3         3         3       3            3           <none>          12m

 

# Daemonset 통해서 rolling update가 가능하다
[master ~]$ kubectl edit daemonsets.apps daemonset-nginx 
daemonset.apps/daemonset-nginx edited

## edit 화면 참고

[master ~]$ kubectl get pods -o wide
NAME                    READY   STATUS              RESTARTS   AGE   IP         NODE    NOMINATED NODE   READINESS GATES
daemonset-nginx-5f8hh   1/1     Running             0          4s    10.5.1.6   node2   <none>           <none>
daemonset-nginx-6q4rq   1/1     Running             0          6s    10.5.2.6   node3   <none>           <none>
daemonset-nginx-t7kwx   0/1     ContainerCreating   0          2s    <none>     node4   <none>           <none>
[master ~]$ kubectl get daemonsets.apps 
NAME              DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
daemonset-nginx   3         3         2       3            2           <none>          13m
[master ~]$ kubectl get daemonsets.apps 
NAME              DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
daemonset-nginx   3         3         3       3            3           <none>          13m
[master ~]$ kubectl get pods -o wide
NAME                    READY   STATUS    RESTARTS   AGE   IP         NODE    NOMINATED NODE   READINESS GATES
daemonset-nginx-5f8hh   1/1     Running   0          16s   10.5.1.6   node2   <none>           <none>
daemonset-nginx-6q4rq   1/1     Running   0          18s   10.5.2.6   node3   <none>           <none>
daemonset-nginx-t7kwx   1/1     Running   0          14s   10.5.4.5   node4   <none>           <none>

[master ~]$ kubectl describe pod daemonset-nginx-5f8hh
Name:             daemonset-nginx-5f8hh
Namespace:        default
Priority:         0
Service Account:  default
Node:             node2/192.168.0.17
Start Time:       Mon, 26 Feb 2024 06:46:39 +0000
Labels:           app=webui
                  controller-revision-hash=6459b7498d
                  pod-template-generation=2
Annotations:      <none>
Status:           Running
IP:               10.5.1.6
IPs:
  IP:           10.5.1.6
Controlled By:  DaemonSet/daemonset-nginx
Containers:
  nginx-container:
    Container ID:   containerd://fed8be33d34d2ebf29596f36a08f9dadfe079702a1eb3406afd6f06bb3c532ad
    Image:          nginx:1.15
    Image ID:       docker.io/library/nginx@sha256:23b4dcdf0d34d4a129755fc6f52e1c6e23bb34ea011b315d87e193033bcd1b68
    Port:           <none>
    Host Port:      <none>
    State:          Running
    
    
# rollback도 가능하다
[master ~]$ kubectl rollout undo daemonset daemonset-nginx 
daemonset.apps/daemonset-nginx rolled back
[master ~]$ kubectl get pods -o wide
NAME                    READY   STATUS        RESTARTS   AGE     IP         NODE    NOMINATED NODE   READINESS GATES
daemonset-nginx-5f8hh   1/1     Running       0          2m50s   10.5.1.6   node2   <none>           <none>
daemonset-nginx-6q4rq   1/1     Terminating   0          2m52s   10.5.2.6   node3   <none>           <none>
daemonset-nginx-mrdnl   1/1     Running       0          2s      10.5.4.6   node4   <none>           <none>
[master ~]$ kubectl get pods -o wide
NAME                    READY   STATUS    RESTARTS   AGE   IP         NODE    NOMINATED NODE   READINESS GATES
daemonset-nginx-gnfgc   1/1     Running   0          4s    10.5.1.7   node2   <none>           <none>
daemonset-nginx-lszzs   1/1     Running   0          6s    10.5.2.7   node3   <none>           <none>
daemonset-nginx-mrdnl   1/1     Running   0          8s    10.5.4.6   node4   <none>           <none>

 

# 실습] > FluentD logging을 위한 DaemonSet Deploy하기

# Deployment를 활용해서 만들기
$ kubectl create deployment elasticsearch -n kube-system --image=registry.k8s.io/fluentd-elasticsearch:1.20 --dry-run=client -o yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  creationTimestamp: null
  labels:
    app: elasticsearch
  name: elasticsearch
  namespace: kube-system
spec:
  replicas: 1
  selector:
    matchLabels:
      app: elasticsearch
  strategy: {}
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: elasticsearch
    spec:
      containers:
      - image: registry.k8s.io/fluentd-elasticsearch:1.20
        name: fluentd-elasticsearch
        resources: {}
status: {}

$ kubectl create deployment elasticsearch -n kube-system --image=registry.k8s.io/fluentd-elasticsearch:1.20 --dry-run=client -o yaml > fluentd.yaml
# DaemonSet에 맞게 수정
$ vi fluentd.yaml
$ cat fluentd.yaml 
apiVersion: apps/v1
kind: DaemonSet
metadata:
  labels:
    app: elasticsearch
  name: elasticsearch
  namespace: kube-system
spec:
  selector:
    matchLabels:
      app: elasticsearch
  template:
    metadata:
      labels:
        app: elasticsearch
    spec:
      containers:
      - image: registry.k8s.io/fluentd-elasticsearch:1.20
        name: fluentd-elasticsearch
$ kubectl create -f fluentd.yaml 
daemonset.apps/elasticsearch created
        
$ kubectl get ds -n kube-system
NAME            DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR            AGE
elasticsearch   1         1         1       1            1           <none>                   116s
kube-proxy      1         1         1       1            1           kubernetes.io/os=linux   48m

 

Rolling Update

DaemonSet은 업데이트 시 Rolling Update 전략을 지원하여 기존 Pod를 순차적으로 교체한다.

참고:  perform a rolling update 

updateStrategy:
  type: RollingUpdate
  rollingUpdate:
    maxUnavailable: 1
    
 # maxUnavailable: 동시에 비활성화될 수 있는 Pod 수(기본값: 1).
 # 기존 Pod를 하나씩 교체하면서 새로운 Pod를 생성.

 

DaemonSet과 Deployment의 차이점

  DaemonSet Deployment
주요 목적 각 노드에 하나의 Pod 실행 보장 지정된 수의 Pod 실행 보장
스케줄링 방식 모든 노드 또는 특정 조건의 노드 노드에 관계없이 스케줄링
사용 사례 로그 수집, 모니터링, 네트워크 에이전트 등 애플리케이션, 서비스 등 일반 워크로드
Pod 개수 노드 수와 동일 사용자가 직접 지정

 

# yaml 생성 시 deployment와 비슷하기 때문에 kubectl create deployment를 활용해서 생성한다.

$  kubectl create deployment elasticsearch -n kube-system --image registry.k8s.io/fluentd-elasticsearch:1.20 --dry-run=client -o yaml > fluentd.yaml
$  vi fluentd.yaml 
apiVersion: apps/v1
kind: DaemonSet
metadata:
  creationTimestamp: null
  labels:
    app: elasticsearch
  name: elasticsearch
  namespace: kube-system
spec:
  selector:
    matchLabels:
      app: elasticsearch
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: elasticsearch
    spec:
      containers:
      - image: registry.k8s.io/fluentd-elasticsearch:1.20
        name: fluentd-elasticsearch
        resources: {}
        
$ kubectl create -f fluentd.yaml 
daemonset.apps/elasticsearch created

$ kubectl get ds -n kube-system 
NAME            DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR            AGE
elasticsearch   1         1         1       1            1           <none>                   28s
kube-proxy      1         1         1       1            1           kubernetes.io/os=linux   19m
반응형

'Container > Kubernetes' 카테고리의 다른 글

[K8S] Job Controller  (0) 2024.02.27
[K8S] Statefulset  (0) 2024.02.26
[K8S] Deployment - RollingUpdate  (0) 2024.02.21
[K8S] ReplicaSet  (0) 2024.02.21
[K8S] ReplicationController  (0) 2024.02.19

'Container/Kubernetes' Related Articles

티스토리툴바