1. 인공 지능과 인공지능 연구의 역사 :


- Environment =? cybernetics(1920-1950)

- Mind(= computer) => Symbolic AI(1950- 1980)

- Brain => Neural Nets(ML) (1980-2010)

- Body=>embodied mind, mind machine => Autonomous Robots(2010~`)


Symbolic AI vs Machine Learning

(기호.논리)       연결성(Neural Network)

                     ML에서 성능을 내기 위한 Source는 Big Data + 좋은 H/W 성능

                     이 필요하다.


*인공지능의 발전 :

Mind: Symbolic AI 

- 1997년 IBM Chess machine Deep Blue

- chatbots(Cornell Univ)


Body : Physical AI

- Humanoid Robot Nao

- Personal Robot PR2(willow Garage)

- Robot roommate shopping for and preparing bavarian breakfast

- Delivery Robots Bots


Embodied Mind

- Smart Speakers


AI가 왜 갑자기 Hot하게 됐는가?
- Machine Learning 알고리즘들이 학자들에 의해 잘 정립이 되었고

- Big Data와 Computing Power(H/W)가 맞물리면서 급속도로 발전하게됨


CNN 기반 얼굴인식 시스템이 지배적이다(?)




2. Deep Learning

모라벡의 역설(Moravec's Paradox)


- 인공지능. 로봇공학 연구자에 따르면 고등 추론에는 연산 능력이 거의 필요 없는 반면,

- 낮은 수준의 감각운동 기능은 엄청난 연산 자원이 필요.

- "어려운 문제는 쉽고 쉬운 문제는 어렵다"


*Deep Learning 의 활용

- Youtube 자동 음석 번역

- Netflix의 영화추천

- Facebook의 feed 추천

- 알파고

- 구글의 검색 결과


*CNN

- 물체들을 추상화하는 과정을 거친다


*RNN :
관련알고리즘 : LSTM, GRU

Image Narration using LSTM

자연어처리에도 많이 사용한다.



저작자표시

'Python 활용 딥러닝' 카테고리의 다른 글

Tensorflow Basic(Ranks, Shapes, Types) + Pycharm 설치  (0) 2018.12.17
Scipy 사용법  (0) 2018.12.17
Numpy 사용법(배열연산, 수학함수)  (0) 2018.12.17
Python if와 for Loop, while문 사용법  (0) 2018.12.17
Python 기본 + Numy, Jupyter 설치 + Python 함수 사용방법  (0) 2018.12.17

*test라는 이름으로 Kubernetes Namespace 생성 및 리소스 사용 : 


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
[root@host01-4 hk]# k create ns test
namespace "test" created
[root@host01-4 hk]# k get ns
NAME          STATUS    AGE
default       Active    6h
kube-public   Active    6h
kube-system   Active    6h
test          Active    4s
[root@host01-4 hk]# ls -l
total 974676
-rw-------. 1 root root  50727424 May 25 10:21 dns-kube-dns
-rw-------. 1 root root  42481152 May 25 10:21 dns-sidecar
-rw-r--r--. 1 root root       158 May 25 16:18 endpoint.yaml
-rw-------. 1 root root 193461760 May 25 10:21 etcd-amd64
-rw-r--r--. 1 root root       362 May 25 12:17 hk.yaml
-rw-------. 1 root root  41239040 May 25 10:21 k8s-dns
-rw-r--r--. 1 root root       343 May 25 10:21 kubeadm.host08-1.root.log.INFO.20180525-060140.2620
-rw-r--r--. 1 root root       343 May 25 10:21 kubeadm.INFO
-rw-------. 1 root root 225319936 May 25 10:21 kube-apiserver
-rw-------. 1 root root 148110336 May 25 10:21 kube-controller
-rw-------. 1 root root  98924032 May 25 10:21 kube-proxy
-rw-------. 1 root root  50635776 May 25 10:21 kube-scheduler
-rw-r--r--. 1 root root       715 May 25 13:45 multi.yaml
-rw-r--r--. 1 root root       185 May 25 16:48 nodePort.yaml
-rw-r--r--. 1 root root       162 May 25 17:03 probe.yaml
-rw-r--r--. 1 root root       410 May 25 17:06 rc-probe.yaml
-rw-r--r--. 1 root root       302 May 25 15:51 rc.yaml
-rw-r--r--. 1 root root         0 May 25 15:53 service.yalm
-rw-r--r--. 1 root root       162 May 25 15:55 service.yaml
-rw-r--r--. 1 root root       322 May 25 14:11 temp.yaml
-rw-r--r--. 1 root root      2336 May 25 12:06 temp.yarm
-rw-r--r--. 1 root root        88 May 25 16:13 test-svc-h1.yaml
-rw-------. 1 root root  99517952 May 25 10:21 weave-kube
-rw-------. 1 root root  47575552 May 25 10:21 weave-npc
[root@host01-4 hk]# k create -f rc.yaml -n test
replicationcontroller "simple-rc" created
[root@host01-4 hk]# k get rc -n test
NAME        DESIRED   CURRENT   READY     AGE
simple-rc   3         3         3         9s
[root@host01-4 hk]#
[root@host01-4 hk]# k describe ns test
 
cs





저작자표시

'Docker(도커) Kubernetes' 카테고리의 다른 글

Docker Kubernetes, Readiness Probe 생성 방법  (0) 2018.05.25
Pod 개념, 멀티 컨테이너 생성, kubectl 명령어 사용법  (0) 2018.05.25
Kubernetes 설치 및 사용법, Yaml으로 pod 및 rc 생성  (0) 2018.05.25
도커 관련 설정 파일 모음  (2) 2018.05.24
docker swarm 사용법  (0) 2018.05.24

LB > Node Port > Cluster IP


Node Port를 만들면 Cluster IP가 자동으로 만들어진다


*SVC 지우기

1
2
3
4
5
6
7
8
9
[root@host01-4 hk]# k get service
NAME           TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)   AGE
headless-svc   ClusterIP   10.96.102.169   <none>        80/TCP    27m
kubernetes     ClusterIP   10.96.0.1       <none>        443/TCP   48m
test-svc       ClusterIP   10.96.25.31     <none>        80/TCP    45m
[root@host01-4 hk]# k delete svc test-svc
service "test-svc" deleted
[root@host01-4 hk]#
 
cs


*NodePort 생성하기(생성뒤에 해당 Port로 curl을 날릴 수 있다) : 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
vi NodePort.yaml
 
apiVersion: v1
kind: Service
metadata:
  name: test-svc
spec:
        type: NodePort
        selector:
          type: test
        ports:
        - port: 80
          targetPort: 8080
 
 
[root@host01-4 hk]# k create -f nodePort.yaml
 
service "test-svc" created
[root@host01-4 hk]#
[root@host01-4 hk]# k get svc
NAME           TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)        AGE
headless-svc   ClusterIP   10.96.102.169   <none>        80/TCP         31m
kubernetes     ClusterIP   10.96.0.1       <none>        443/TCP        52m
test-svc       NodePort    10.107.213.48   <none>        80:32045/TCP   6s
 
 
#NodePort에서 80포트는 cluster IP port이고 32045는 Node Port
 
 
cs


*Load Balancer 생성

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
k edit svc test-app
 
# Please edit the object below. Lines beginning with a '#' will be ignored,
# and an empty file will abort the edit. If an error occurs while saving this file will be
# reopened with the relevant failures.
#
apiVersion: v1
kind: Service
metadata:
  creationTimestamp: 2018-05-25T07:44:57Z
  name: test-svc
  namespace: default
  resourceVersion: "29844"
  selfLink: /api/v1/namespaces/default/services/test-svc
  uid: 85b2c6ea-5fef-11e8-8e09-005056b28b62
spec:
  clusterIP: 10.107.213.48
  externalTrafficPolicy: Cluster
  ports:
  - nodePort: 32045
    port: 80
    protocol: TCP
    targetPort: 8080
  selector:
    type: test
  sessionAffinity: None
  type: LoadBalancer            #LoadBalancer로 수정해주면...
status:
  loadBalancer: {}
 
 
[root@host01-4 hk]# k get svc
NAME           TYPE           CLUSTER-IP      EXTERNAL-IP   PORT(S)        AGE
headless-svc   ClusterIP      10.96.102.169   <none>        80/TCP         38m
kubernetes     ClusterIP      10.96.0.1       <none>        443/TCP        59m
test-svc       LoadBalancer   10.107.213.48   <pending>     80:32045/TCP   7m  #Node Port에 Cluster IP까지 자동 
[root@host01-4 hk]#
 
cs


상위 네트워크 모듈을 생성하면 하위 매핑 정보는 알아서 구성하게 됨




*TargetPort 설정 이유


1
2
3
4
5
6
7
 
  ports:
  - nodePort: 32045
    targetPort: 8080
 
 
 
cs


=> Cluster의 port와 pod Port 정보가 다르기 때문



*비즈니스 관점에서 Probe가 정상적으로 동작하는지 여부를 체크할 수 있어야 한다....


Client가 접속한다!!

Svc > Pod(X)

정상동작 X 다고 생각하게 된다.

이때 readinessProbe를 활용하면 cluster 정상동작 여부를 체크할 수 있다 :

1
2
3
4
5
6
7
8
9
      containers:
      - image: reg.cloud.com/kubia
        name: kubia
        readinessProbe:
          exec:
            command:
            - ls
            - /var/ready
 
cs


*ReadinessProbe 생성 예제 :


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
#vi rc-probe.yaml 
# /var/ready가 pod 안에 있으면 probe에서 통과 시킨다(Ready 상태로 만들어줌)
apiVersion: v1
kind: ReplicationController
metadata:
  name: rc-readiness
spec:
  replicas: 3
  template:
    metadata:
      labels:
        type: test
    spec:
      containers:
      - image: reg.cloud.com/kubia
        name: kubia
        readinessProbe:
          exec:
            command:
            - ls
            - /var/ready
        ports:
        - containerPort: 8080
          protocol: TCP
 
 
#vi service.yaml
 
apiVersion: v1
kind: Service
metadata:
  name: test-svc
spec:
        selector:
          type: test
        ports:
        - port: 80
          targetPort: 8080
 
 
 
[root@host01-4 hk]# k create -f rc-probe.yaml
 
replicationcontroller "rc-readiness" created
[root@host01-4 hk]# k get po -o wide
NAME                 READY     STATUS    RESTARTS   AGE       IP          NODE
rc-readiness-22hxs   0/1       Running   0          10s       10.36.0.2   host01-3.cloud.com
rc-readiness-2pt29   0/1       Running   0          10s       10.44.0.1   host01-2.cloud.com
rc-readiness-h74t2   0/1       Running   0          10s       10.36.0.1   host01-3.cloud.com
 
 
#Running 상태이지만 Ready가 아무것도 없다
 
 
 
[root@host01-4 hk]# k create -f service.yaml
service "test-svc" created
[root@host01-4 hk]# k get svc
NAME         TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)   AGE
kubernetes   ClusterIP   10.96.0.1     <none>        443/TCP   6m
test-svc     ClusterIP   10.98.99.22   <none>        80/TCP    4s
[root@host01-4 hk]# k get ep
NAME         ENDPOINTS          AGE
kubernetes   10.10.12.14:6443   6m
test-svc                        9s
 
 
#Endpoint가 아무것도 안 떠있다.
 
[root@host01-4 hk]# k exec rc-readiness-22hxs -- touch /var/ready
 
#1번 pod에 /var/ready를 만들어준다 (Ready 상태로 만들어주기 위해)
 
[root@host01-4 hk]# k get ep
NAME         ENDPOINTS          AGE
kubernetes   10.10.12.14:6443   8m
test-svc     10.36.0.2:8080     1m
 
#test-svc의 ENDPOINT가 enabled된다
 
[root@host01-4 hk]# k get svc
NAME         TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)   AGE
kubernetes   ClusterIP   10.96.0.1     <none>        443/TCP   8m
test-svc     ClusterIP   10.98.99.22   <none>        80/TCP    1m
 
[root@host01-4 hk]# k get po -o wide
NAME                 READY     STATUS    RESTARTS   AGE       IP          NODE
rc-readiness-22hxs   1/1       Running   0          4m        10.36.0.2   host01-3.cloud.com
rc-readiness-2pt29   0/1       Running   0          4m        10.44.0.1   host01-2.cloud.com
rc-readiness-h74t2   0/1       Running   0          4m        10.36.0.1   host01-3.cloud.com
 
#1번 Pod가 Ready 상태로 바뀐다
 
 
[root@host01-4 hk]# k exec rc-readiness-2pt29 -- touch /var/ready
[root@host01-4 hk]# k get po -o wide
NAME                 READY     STATUS    RESTARTS   AGE       IP          NODE
rc-readiness-22hxs   1/1       Running   0          5m        10.36.0.2   host01-3.cloud.com
rc-readiness-2pt29   1/1       Running   0          5m        10.44.0.1   host01-2.cloud.com
rc-readiness-h74t2   0/1       Running   0          5m        10.36.0.1   host01-3.cloud.com
 
#2번 Pod Ready 상태로 바뀐다
 
 
 
 
cs






저작자표시

'Docker(도커) Kubernetes' 카테고리의 다른 글

Kubernetes Namespace 생성 및 리소스 사용  (0) 2018.05.25
Pod 개념, 멀티 컨테이너 생성, kubectl 명령어 사용법  (0) 2018.05.25
Kubernetes 설치 및 사용법, Yaml으로 pod 및 rc 생성  (0) 2018.05.25
도커 관련 설정 파일 모음  (2) 2018.05.24
docker swarm 사용법  (0) 2018.05.24

*Pod = 컨테이너 그룹

rc로 제어되는 pod는 삭제하면 다시 생성된다

반면 자체적으로 만든 pod는 삭제하면 재생성되지 않음



*멀티 컨테이너 :

- create pod

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
multi.yaml
 
apiVersion: v1
kind: Pod
metadata:
  name: multi
  labels:
    app: wordpress
spec:
  containers:
    - name: wordpress
      image: reg.cloud.com/wordpress
      ports:
        - containerPort: 80
        - containerPort: 443
      env:
        - name: WORDPRESS_DB_HOST
          value: 127.0.0.1
        - name: WORDPRESS_DB_USER
          value: wordpress
        - name: WORDPRESS_DB_PASSWORD
          value: password
    - name: db
      image: reg.cloud.com/mysql
      env:
        - name: MYSQL_ROOT_PASSWORD
          value: password
        - name: MYSQL_DATABASE
          value: wordpress
        - name: MYSQL_USER
          value: wordpress
        - name: MYSQL_PASSWORD
          value: password
 
 
[root@host01-4 hk]# k create -f multi.yaml
 
cs



- 멀티 컨테이너에서는 컨테이너를 지정해서 들어갈 수 있다(DB랑 web이 하나의 포드로 묶여있기 때문에 좋은 구성은 아니다).

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
[root@host01-4 hk]# k get po -o wide
NAME               READY     STATUS    RESTARTS   AGE       IP          NODE
multi              2/2       Running   1          1m        10.44.0.2   host01-2.cloud.com
nginx-app-7qpbv    1/1       Running   0          1h        10.36.0.1   host01-3.cloud.com
nginx-app2-tgqqf   1/1       Running   0          1h        10.36.0.2   host01-3.cloud.com
nginx-hk-app       1/1       Running   0          1h        10.44.0.1   host01-2.cloud.com
[root@host01-4 hk]# k exec -it multi -c wordpress sh
# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
18: eth0@if19: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1376 qdisc noqueue state UP group default
    link/ether 32:a0:1d:fb:91:b1 brd ff:ff:ff:ff:ff:ff
    inet 10.44.0.2/12 brd 10.47.255.255 scope global eth0
       valid_lft forever preferred_lft forever
    inet6 fe80::30a0:1dff:fefb:91b1/64 scope link tentative dadfailed
       valid_lft forever preferred_lft forever
# ^C
# hostname
multi
# [root@host01-4 hk]# k exec -it multi -c db sh
# mysql
ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: NO)
# mysql -uroot -p
Enter password:
ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: YES)
#
 
cs





*Label 정보 확인

1
2
3
4
5
6
7
8
9
 
[root@host01-4 hk]# k get po --show-labels
NAME               READY     STATUS    RESTARTS   AGE       LABELS
multi              2/2       Running   1          5m        app=wordpress
nginx-app-7qpbv    1/1       Running   0          1h        run=nginx-app
nginx-app2-tgqqf   1/1       Running   0          1h        type=test
nginx-hk-app       1/1       Running   0          1h        type=web
[root@host01-4 hk]#
 
cs


*Label 조회/생성/Overwrite

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
[root@host01-4 hk]# k get po --show-labels
NAME               READY     STATUS    RESTARTS   AGE       LABELS
multi              2/2       Running   1          5m        app=wordpress
nginx-app-7qpbv    1/1       Running   0          1h        run=nginx-app
nginx-app2-tgqqf   1/1       Running   0          1h        type=test
nginx-hk-app       1/1       Running   0          1h        type=web
[root@host01-4 hk]# ^C
[root@host01-4 hk]# k label po simple-pod loc=seoul
Error from server (NotFound): pods "simple-pod" not found
[root@host01-4 hk]# k label po nginx-app-7qpbv loc=seoul
pod "nginx-app-7qpbv" labeled
[root@host01-4 hk]# k label po nginx-app-7qpbv type=was --overwrite
pod "nginx-app-7qpbv" labeled
[root@host01-4 hk]#
 
cs



*Pod 찾기(-l 옵션으로 key=value 방식으로 가져올 수 있음) :

1
2
3
4
5
6
7
8
9
10
11
[root@host01-4 hk]# k get po -l type=test
NAME               READY     STATUS    RESTARTS   AGE
nginx-app2-tgqqf   1/1       Running   0          1h
[root@host01-4 hk]# ^C
[root@host01-4 hk]# k get po -l type
NAME               READY     STATUS    RESTARTS   AGE
nginx-app-7qpbv    1/1       Running   0          1h
nginx-app2-tgqqf   1/1       Running   0          1h
nginx-hk-app       1/1       Running   0          1h
[root@host01-4 hk]#
 
cs


*Pod 찾기(-L : Label로 찾기)

1
2
3
4
5
6
7
8
[root@host01-4 hk]# k get po -L app,type,loc
NAME               READY     STATUS    RESTARTS   AGE       APP         TYPE      LOC
multi              2/2       Running   1          11m       wordpress
nginx-app-7qpbv    1/1       Running   0          1h                    was       seoul
nginx-app2-tgqqf   1/1       Running   0          1h                    test
nginx-hk-app       1/1       Running   0          1h                    web
[root@host01-4 hk]#
 
cs



1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
[root@host01-4 hk]# k get po
NAME               READY     STATUS    RESTARTS   AGE
multi              2/2       Running   1          15m
nginx-app-7qpbv    1/1       Running   0          1h
nginx-app2-tgqqf   1/1       Running   0          1h
nginx-hk-app       1/1       Running   0          1h
nginx-hk-app2      0/1       Pending   0          32s
[root@host01-4 hk]# k get no
NAME                 STATUS    ROLES     AGE       VERSION
host01-2.cloud.com   Ready     <none>    3h        v1.10.3
host01-3.cloud.com   Ready     <none>    3h        v1.10.3
host01-4.cloud.com   Ready     master    3h        v1.10.3
 
[root@host01-4 hk]# k label no host01-2.cloud.com gpu=true
node "host01-2.cloud.com" labeled
[root@host01-4 hk]# k get no --show-labels
NAME                 STATUS    ROLES     AGE       VERSION   LABELS
host01-2.cloud.com   Ready     <none>    3h        v1.10.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,gpu=true,kubernetes.io/hostname=host01-2.cloud.com
host01-3.cloud.com   Ready     <none>    3h        v1.10.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/hostname=host01-3.cloud.com
host01-4.cloud.com   Ready     master    3h        v1.10.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/hostname=host01-4.cloud.com,node-role.kubernetes.io/master=
[root@host01-4 hk]# k get po -o wide
NAME               READY     STATUS    RESTARTS   AGE       IP          NODE
multi              2/2       Running   1          17m       10.44.0.2   host01-2.cloud.com
nginx-app-7qpbv    1/1       Running   0          1h        10.36.0.1   host01-3.cloud.com
nginx-app2-tgqqf   1/1       Running   0          1h        10.36.0.2   host01-3.cloud.com
nginx-hk-app       1/1       Running   0          1h        10.44.0.1   host01-2.cloud.com
nginx-hk-app2      1/1       Running   0          3m        10.44.0.3   host01-2.cloud.com
 
cs





* 전체 삭제

1
2
3
4
5
6
7
[root@host01-4 hk]# k delete all --all
pod "multi" deleted
service "kubernetes" deleted
[root@host01-4 hk]# k get rc
No resources found.
[root@host01-4 hk]#
 
cs


*create rc(2replicas) + describe

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
apiVersion: v1
kind: ReplicationController
metadata:
  labels:
  name: simple-rc
spec:
  replicas: 2
  template:
    metadata:
      labels:
        type: test
    spec:
      containers:
      - image: reg.cloud.com/nginx
        name: nginx-app
        ports:
        - containerPort: 80
          protocol: TCP
 
 
[root@host01-4 hk]# k create -f rc.yaml
replicationcontroller "simple-rc" created
[root@host01-4 hk]# get rc
-bash: get: command not found
[root@host01-4 hk]# k get rc
NAME        DESIRED   CURRENT   READY     AGE
simple-rc   2         2         2         5s
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
simple-rc-26m58   1/1       Running   0          21s       10.44.0.1   host01-2.cloud.com
simple-rc-7htdc   1/1       Running   0          21s       10.36.0.1   host01-3.cloud.com
[root@host01-4 hk]# vi rc.yaml
 
 
[root@host01-4 hk]# k describe rc simple-rc
Name:         simple-rc
Namespace:    default
Selector:     type=test
Labels:       type=test
Annotations:  <none>
Replicas:     2 current / 2 desired
Pods Status:  2 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:  type=test
  Containers:
   nginx-app:
    Image:        reg.cloud.com/nginx
    Port:         80/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:       <none>
  Volumes:        <none>
Events:
  Type    Reason            Age   From                    Message
  ----    ------            ----  ----                    -------
  Normal  SuccessfulCreate  1m    replication-controller  Created pod: simple-rc-7htdc
  Normal  SuccessfulCreate  1m    replication-controller  Created pod: simple-rc-26m58
[root@host01-4 hk]#
 
cs



*Label 조회

1
2
3
4
5
6
[root@host01-4 hk]# k get po --show-labels
NAME              READY     STATUS    RESTARTS   AGE       LABELS
simple-rc-26m58   1/1       Running   0          2m        type=test
simple-rc-7htdc   1/1       Running   0          2m        type=test
[root@host01-4 hk]#
 
cs


*Label 값 수정 : 

1
2
3
4
5
6
7
8
root@host01-4 hk]# k label po simple-rc-26m58 loc=se
pod "simple-rc-26m58" labeled
[root@host01-4 hk]# k get po --show-labels
NAME              READY     STATUS    RESTARTS   AGE       LABELS
simple-rc-26m58   1/1       Running   0          3m        loc=se,type=test
simple-rc-7htdc   1/1       Running   0          3m        type=test
[root@host01-4 hk]#
 
cs


*Label overwrite(하나의 pod가 추가 생성된다)

1
2
3
4
5
6
7
8
9
10
11
12
13
[root@host01-4 hk]# k get po --show-labels
NAME              READY     STATUS    RESTARTS   AGE       LABELS
simple-rc-26m58   1/1       Running   0          4m        loc=se,type=test
simple-rc-7htdc   1/1       Running   0          4m        type=test
[root@host01-4 hk]# k label po simple-rc-26m58 type=op --overwrite
pod "simple-rc-26m58" labeled
[root@host01-4 hk]# k get po --show-labels
NAME              READY     STATUS    RESTARTS   AGE       LABELS
simple-rc-26m58   1/1       Running   0          5m        loc=se,type=op
simple-rc-7htdc   1/1       Running   0          5m        type=test
simple-rc-bts7r   1/1       Running   0          3s        type=test
[root@host01-4 hk]#
 
cs


*RC replicas Scale out :

1
2
3
[root@host01-4 hk]# k scale rc simple-rc --replicas=3
replicationcontroller "simple-rc" scaled
 
cs


*RC를 yaml 파일로 띄워서 수정(replicas부분을 2로 변경 시 바로 2개로 수정되어 뜬다- 1가 죽음)

1
2
3
4
5
6
7
8
9
[root@host01-4 hk]# k edit rc simple-rc
replicationcontroller "simple-rc" edited
[root@host01-4 hk]# k get po --show-labels
NAME              READY     STATUS        RESTARTS   AGE       LABELS
simple-rc-26m58   1/1       Running       0          8m        loc=se,type=op
simple-rc-7htdc   1/1       Running       0          8m        type=test
simple-rc-9cmhk   0/1       Terminating   0          51s       type=test
simple-rc-bts7r   1/1       Running       0          3m        type=test
 
cs


*disk를 ssd로 해서 띄우기

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
apiVersion: apps/v1beta2
kind: DaemonSet
metadata:
  name: simple-ds
spec:
  selector:
    matchLabels:
      type: app #selector는 template.labels.type과 
  template:
    metadata:
      labels:
        type: app
    spec:
      nodeSelector:
        disk: ssd  # 추가됨
      containers:
      - image: reg.cloud.com/nginx
        name: nginx-app
        ports:
        - containerPort: 80
          protocol: TCP
 
 
[root@host01-4 hk]# vi rc.yaml
[root@host01-4 hk]# k create -f rc.yaml
daemonset.apps "simple-ds" created
[root@host01-4 hk]#
 
cs


*node 조회

1
2
3
4
5
6
[root@host01-4 hk]# k get no
NAME                 STATUS    ROLES     AGE       VERSION
host01-2.cloud.com   Ready     <none>    4h        v1.10.3
host01-3.cloud.com   Ready     <none>    4h        v1.10.3
host01-4.cloud.com   Ready     master    4h        v1.10.3
 
cs


*Labeling 후 pod로 ssd가 생성이 되었는지 확인(disk=ssd)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
[root@host01-4 hk]# k get no
NAME                 STATUS    ROLES     AGE       VERSION
host01-2.cloud.com   Ready     <none>    4h        v1.10.3
host01-3.cloud.com   Ready     <none>    4h        v1.10.3
host01-4.cloud.com   Ready     master    4h        v1.10.3
[root@host01-4 hk]# k label no host01-2.cloud.com disk=ssd
node "host01-2.cloud.com" labeled
[root@host01-4 hk]# get no --show-labels
-bash: get: command not found
[root@host01-4 hk]# k get no --show-labels
NAME                 STATUS    ROLES     AGE       VERSION   LABELS
host01-2.cloud.com   Ready     <none>    4h        v1.10.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,disk=ssd,gpu=true,kubernetes.io/hostname=host01-2.cloud.com
host01-3.cloud.com   Ready     <none>    4h        v1.10.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/hostname=host01-3.cloud.com
host01-4.cloud.com   Ready     master    4h        v1.10.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/hostname=host01-4.cloud.com,node-role.kubernetes.io/master=
[root@host01-4 hk]# k get ds
NAME        DESIRED   CURRENT   READY     UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
simple-ds   1         1         1         1            1           disk=ssd        3m
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
simple-ds-vv5lp   1/1       Running   0          1m        10.44.0.3   host01-2.cloud.com
simple-rc-26m58   1/1       Running   0          22m       10.44.0.1   host01-2.cloud.com
simple-rc-7htdc   1/1       Running   0          22m       10.36.0.1   host01-3.cloud.com
simple-rc-bts7r   1/1       Running   0          17m       10.44.0.2   host01-2.cloud.com
[root@host01-4 hk]#
 
cs



*Labeling을 disk=hhd로 overwrite 할 시 다시 pod 리스트에서 빠져버린다.


1
2
3
4
[root@host01-4 hk]# k get ds
NAME        DESIRED   CURRENT   READY     UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
simple-ds   0         0         0         0            0           disk=ssd        8m
 
cs

1
2
3
4
5
6
7
8
9
[root@host01-4 hk]# k label no host01-2.cloud.com disk=hdd --overwrite
node "host01-2.cloud.com" labeled
 
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
simple-rc-26m58   1/1       Running   0          23m       10.44.0.1   host01-2.cloud.com
simple-rc-7htdc   1/1       Running   0          23m       10.36.0.1   host01-3.cloud.com
simple-rc-bts7r   1/1       Running   0          18m       10.44.0.2   host01-2.cloud.com
 
cs

글




*내부 연결에 있어서는 Cluster IP가 Core Service!!!

(External LB > Node Port(Kubernetes) > Cluster IP(Kubernetes))


네임으로 날리면 DNS에서 제어해서 

알아서 백단의 VIP로 redirection 시켜준다


1) External LB => Node Port=> DB(NS) => VIP => DB POD

2) External LB => Node Port=> LB(NS) => VIP => WEB POD1, WEB POD2



*kubectl pod Watch 명령어 :

1
2
3
4
5
[root@host01-4 hk]# k get po -o wide --watch
NAME              READY     STATUS    RESTARTS   AGE       IP        NODE
simple-rc-bdvs6   0/1       Pending   0          38s       <none>    <none>
simple-rc-d5h74   0/1       Pending   0          38s       <none>    <none>
simple-rc-fq6hl   0/1       Pending   0          38s       <none>    <none>
cs

 



*Service 만들기(Cluster IP = 내부 LB역할을 한다) - Pod IP는 가변적으로 기억할 필요 없음 :


-service 조회

1
2
3
4
5
[root@host01-4 hk]# k get svc
NAME         TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)   AGE
kubernetes   ClusterIP   10.96.0.1     <none>        443/TCP   12m
test-svc     ClusterIP   10.96.25.31   <none>        80/TCP    8m
 
cs


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
vi service yalm
 
apiVersion: v1
kind: Service
metadata:
  name: test-svc
spec:
        selector:
          type: test
        ports:
        - port: 80
          targetPort: 8080
 
[root@host01-4 hk]# k create -f service.yaml
service "test-svc" created
 
[root@host01-4 hk]# k get service
NAME         TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)   AGE
kubernetes   ClusterIP   10.96.0.1     <none>        443/TCP   3m
test-svc     ClusterIP   10.96.25.31   <none>        80/TCP    6s
 
 
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
simple-rc-c8bcs   1/1       Running   0          4m        10.36.0.1   host01-3.cloud.com
simple-rc-nfssk   1/1       Running   0          4m        10.44.0.1   host01-2.cloud.com
simple-rc-zgcjv   1/1       Running   0          4m        10.36.0.2   host01-3.cloud.com
[root@host01-4 hk]# curl 10.96.25.31
You've hit simple-rc-zgcjv
[root@host01-4 hk]# curl 10.96.25.31
You've hit simple-rc-c8bcs
[root@host01-4 hk]# curl 10.96.25.31
You've hit simple-rc-zgcjv
cs


*Service 이름 만으로 Curl 가능

1
2
3
4
5
6
7
8
9
10
11
12
13
14
[root@host01-4 hk]# k exec -it simple-rc-c8bcs bash
root@simple-rc-c8bcs:/# curl http://test-svc
 
You've hit simple-rc-c8bcs
root@simple-rc-c8bcs:/#
root@simple-rc-c8bcs:/# cat /etc/resolv.conf
nameserver 10.96.0.10
search default.svc.cluster.local svc.cluster.local cluster.local cloud.com
options ndots:5
root@simple-rc-c8bcs:/# curl http://test-svc.default
You've hit simple-rc-zgcjv
root@simple-rc-c8bcs:/# curl http://test-svc.default.svc
You've hit simple-rc-zgcjv
cs


*한번 연결되면 동일 Client는 동일 IP로 가도록 설정(similar to Sticky Session) : 


1
2
sessionAffinity: ClientIP  #  부분을 ClientIP로 바꿔주면 Sticky Session으로 변경된다
 
cs


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
 
 
 
[root@host01-4 hk]# k edit svc test-svc
service "test-svc" edited
 
# Please edit the object below. Lines beginning with a '#' will be ignored,
# and an empty file will abort the edit. If an error occurs while saving this file will be
# reopened with the relevant failures.
#
apiVersion: v1
kind: Service
metadata:
  creationTimestamp: 2018-05-25T06:55:52Z
  name: test-svc
  namespace: default
  resourceVersion: "25644"
  selfLink: /api/v1/namespaces/default/services/test-svc
  uid: aa440afa-5fe8-11e8-8e09-005056b28b62
spec:
  clusterIP: 10.96.25.31
  ports:
  - port: 80
    protocol: TCP
    targetPort: 8080
  selector:
    type: test
  sessionAffinity: ClientIP  #  부분을 ClientIP로 바꿔주면 Sticky Session으로 변경된다
  type: ClusterIP
status:
  loadBalancer: {}
 
 
[root@host01-4 hk]# k get svc
NAME         TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)   AGE
kubernetes   ClusterIP   10.96.0.1     <none>        443/TCP   15m
test-svc     ClusterIP   10.96.25.31   <none>        80/TCP    11m
 
[root@host01-4 hk]# k get pod -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
simple-rc-c8bcs   1/1       Running   0          15m       10.36.0.1   host01-3.cloud.com
simple-rc-nfssk   1/1       Running   0          15m       10.44.0.1   host01-2.cloud.com
simple-rc-zgcjv   1/1       Running   0          15m       10.36.0.2   host01-3.cloud.com
 
[root@host01-4 hk]# curl 10.96.25.31
You've hit simple-rc-nfssk
[root@host01-4 hk]# curl 10.96.25.31
You've hit simple-rc-nfssk
[root@host01-4 hk]# curl 10.96.25.31
You've hit simple-rc-nfssk
[root@host01-4 hk]#
cs



*Service Endpoint 확인(매핑 정보를 알 수 있다) :

1
2
3
4
5
[root@host01-4 hk]# k get ep
NAME         ENDPOINTS                                      AGE
kubernetes   10.10.12.14:6443                               17m
test-svc     10.36.0.1:8080,10.36.0.2:8080,10.44.0.1:8080   13m
 
cs



*Endpoint 만들기


endpoint는 service와 이름(label name)이 같아야 한다:

> k get svc 

입력 시 뜨는 name과 동일...그 이후에 

> k get ep 

를 입력 시 endpoint로 ip port가 잘 매핑되어 출력되는 것을 확인 할 수 있다



1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
[root@host01-4 hk]# k get svc
NAME           TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)   AGE
headless-svc   ClusterIP   10.96.102.169   <none>        80/TCP    2m
kubernetes     ClusterIP   10.96.0.1       <none>        443/TCP   23m
test-svc       ClusterIP   10.96.25.31     <none>        80/TCP    19m
 
vi endpoint.yaml
 
apiVersion: v1
kind: Endpoints
metadata:
  name: headless-svc
subsets:
  - addresses:
    - ip: 10.10.10.10
    - ip: 11.11.11.11
    ports:
      - port: 80
 
 
#위에서 metadata.name = service name(headless-svc )과 같아야 한다
 
 
 
[root@host01-4 hk]# k create -f endpoint.yaml
endpoints "headless-svc" created
[root@host01-4 hk]# k get service
NAME           TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)   AGE
headless-svc   ClusterIP   10.96.102.169   <none>        80/TCP    6m
kubernetes     ClusterIP   10.96.0.1       <none>        443/TCP   27m
test-svc       ClusterIP   10.96.25.31     <none>        80/TCP    23m
 
[root@host01-4 hk]# k get ep
NAME           ENDPOINTS                                      AGE
headless-svc   10.10.10.10:80,11.11.11.11:80                  19s
kubernetes     10.10.12.14:6443                               27m
test-svc       10.36.0.1:8080,10.36.0.2:8080,10.44.0.1:8080   24m
 
cs

저작자표시

'Docker(도커) Kubernetes' 카테고리의 다른 글

Kubernetes Namespace 생성 및 리소스 사용  (0) 2018.05.25
Docker Kubernetes, Readiness Probe 생성 방법  (0) 2018.05.25
Kubernetes 설치 및 사용법, Yaml으로 pod 및 rc 생성  (0) 2018.05.25
도커 관련 설정 파일 모음  (2) 2018.05.24
docker swarm 사용법  (0) 2018.05.24

*Kubernetes는 다음을 자동으로 처리하는 common API와 self-heading 프레임워크를 제공한다 : 

- 머신 장애

- 어플리케이션 배포, 로깅 모니터링 간소화


최소한의 통제로 자율성을 얻기 위한 툴이다!


점차 Operation 팀의 역할은 줄어들고 있다...

비즈니스는 점점 서비스와 어플리케이션에 포커스를 두고...인프라는 비중이 작아지고 있음


*Kubernetes의 구성도 : 


1) API Node

Kubernetes의 모든 API서버는 Node들의 모든 정보를 가지고 있어야 한다.

스케줄러가 제일 먼저 하는 일은....어디에 배포할 것인가 결정

API Node 내에 controller manager(replication controller)는 node의 status를 알려준다.

->Kubernetes는 ETCD라는 분산형태 DB를 사용한다,



2) Worker Node 

Core OS가 Rocket(docker가 너무 무거워)이라는 컨테이너를 발표!

Kubelet은 컨테이너에 대한 구성요소 제어를 맡고 있음

각 pod간 통신은 VXLAN(overlay)를 사용

proxy는 app 레벨의 인스턴스가 있고,

실제 호스트에 있는 iptables (kernel 레이어에 있음)에 데이터 처리에 관한 내용이 있음



※ kubelet은 유일하게 시스템 서비스로 들어간다! 나머지는 컨테이너로 설치 가능



- 왜 Swarm을 사용하는가? 사용하기 쉬움

- 왜 Kubernetes를 사용하는가? 많은 Resource를 제어할 수 있음





1. Docker 설치 (모든 노드)


1
2
3
4
5
6
7
8
9
yum install -y docker
systemctl enable docker && systemctl start docker
 
touch /etc/docker/daemon.json
cat <<EOF > /etc/docker/daemon.json
{
"insecure-registries":["10.10.12.0/24"]
}
EOF
cs


2. Kubernetes 설치 (모든 노드)


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=https://packages.cloud.google.com/yum/repos/kubernetes-el7-x86_64
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg
EOF
setenforce 0
sed -i --follow-symlinks 's/SELINUX=enforcing/SELINUX=disabled/g' /etc/sysconfig/selinux
systemctl stop firewalld.service
systemctl disable firewalld.service
 
cat <<EOF >  /etc/sysctl.d/k8s.conf
net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
EOF
sysctl --system
 
swapoff -a && sed -i '/swap/s/^/#/' /etc/fstab
 
yum install -y kubelet kubeadm kubectl
systemctl enable kubelet && systemctl start kubelet


*MASTER Node 설치(이미지를 전부 다 Load한다) :

> docker load -i <이미지명>

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
[root@host01-2 ~]# docker images
REPOSITORY          TAG                 IMAGE ID            CREATED             SIZE
[root@host01-2 ~]# ls
anaconda-ks.cfg  dns-kube-dns  dns-sidecar  etcd-amd64  hk
[root@host01-2 ~]# cd hk
[root@host01-2 hk]# ls
dns-kube-dns  etcd-amd64  kubeadm.host08-1.root.log.INFO.20180525-060140.2620  kube-apiserver   kube-proxy      weave-kube
dns-sidecar   k8s-dns     kubeadm.INFO                                         kube-controller  kube-scheduler  weave-npc
[root@host01-2 hk]# ls -al
total 974632
drwxr-xr-x. 2 root root       277 May 25 10:27 .
dr-xr-x---. 5 root root       226 May 25 10:27 ..
-rw-------. 1 root root  50727424 May 25 10:27 dns-kube-dns
-rw-------. 1 root root  42481152 May 25 10:27 dns-sidecar
-rw-------. 1 root root 193461760 May 25 10:27 etcd-amd64
-rw-------. 1 root root  41239040 May 25 10:27 k8s-dns
-rw-r--r--. 1 root root       343 May 25 10:27 kubeadm.host08-1.root.log.INFO.20180525-060140.2620
-rw-r--r--. 1 root root       343 May 25 10:27 kubeadm.INFO
-rw-------. 1 root root 225319936 May 25 10:27 kube-apiserver
-rw-------. 1 root root 148110336 May 25 10:27 kube-controller
-rw-------. 1 root root  98924032 May 25 10:27 kube-proxy
-rw-------. 1 root root  50635776 May 25 10:27 kube-scheduler
-rw-------. 1 root root  99517952 May 25 10:27 weave-kube
-rw-------. 1 root root  47575552 May 25 10:27 weave-npc
[root@host01-2 hk]# docker load -i kube-proxy
582b548209e1: Loading layer [==================================================>]  44.2 MB/44.2 MB
e20569a478ed: Loading layer [==================================================>] 3.358 MB/3.358 MB
6b4e4941a965: Loading layer [==================================================>] 51.35 MB/51.35 MB
Loaded image: k8s.gcr.io/kube-proxy-amd64:v1.10.3
[root@host01-2 hk]# docker load -i weave-kube
5bef08742407: Loading layer [==================================================>] 4.221 MB/4.221 MB
c3355c8b5c3e: Loading layer [==================================================>] 19.03 MB/19.03 MB
a83fa3df4138: Loading layer [==================================================>] 29.55 MB/29.55 MB
020fdc01af85: Loading layer [==================================================>]  11.6 MB/11.6 MB
2ea881a632b7: Loading layer [==================================================>] 2.048 kB/2.048 kB
396aa46bcbea: Loading layer [==================================================>] 35.09 MB/35.09 MB
Loaded image: docker.io/weaveworks/weave-kube:2.3.0
[root@host01-2 hk]# docker load -i weave-npc
8dccfe2dec8c: Loading layer [==================================================>] 2.811 MB/2.811 MB
3249ff6df12f: Loading layer [==================================================>] 40.52 MB/40.52 MB
3dc458d34b22: Loading layer [==================================================>]  2.56 kB/2.56 kB
Loaded image: docker.io/weaveworks/weave-npc:2.3.0
[root@host01-2 hk]#
 
cs

*기타 Node 설치(kube-proxy, weave-kube, weave-npc를 설치한다) :

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
[root@host01-2 ~]# docker images
REPOSITORY          TAG                 IMAGE ID            CREATED             SIZE
[root@host01-2 ~]# ls
anaconda-ks.cfg  dns-kube-dns  dns-sidecar  etcd-amd64  hk
[root@host01-2 ~]# cd hk
[root@host01-2 hk]# ls
dns-kube-dns  etcd-amd64  kubeadm.host08-1.root.log.INFO.20180525-060140.2620  kube-apiserver   kube-proxy      weave-kube
dns-sidecar   k8s-dns     kubeadm.INFO                                         kube-controller  kube-scheduler  weave-npc
[root@host01-2 hk]# ls -al
total 974632
drwxr-xr-x. 2 root root       277 May 25 10:27 .
dr-xr-x---. 5 root root       226 May 25 10:27 ..
-rw-------. 1 root root  50727424 May 25 10:27 dns-kube-dns
-rw-------. 1 root root  42481152 May 25 10:27 dns-sidecar
-rw-------. 1 root root 193461760 May 25 10:27 etcd-amd64
-rw-------. 1 root root  41239040 May 25 10:27 k8s-dns
-rw-r--r--. 1 root root       343 May 25 10:27 kubeadm.host08-1.root.log.INFO.20180525-060140.2620
-rw-r--r--. 1 root root       343 May 25 10:27 kubeadm.INFO
-rw-------. 1 root root 225319936 May 25 10:27 kube-apiserver
-rw-------. 1 root root 148110336 May 25 10:27 kube-controller
-rw-------. 1 root root  98924032 May 25 10:27 kube-proxy
-rw-------. 1 root root  50635776 May 25 10:27 kube-scheduler
-rw-------. 1 root root  99517952 May 25 10:27 weave-kube
-rw-------. 1 root root  47575552 May 25 10:27 weave-npc
[root@host01-2 hk]# docker load -i kube-proxy
582b548209e1: Loading layer [==================================================>]  44.2 MB/44.2 MB
e20569a478ed: Loading layer [==================================================>] 3.358 MB/3.358 MB
6b4e4941a965: Loading layer [==================================================>] 51.35 MB/51.35 MB
Loaded image: k8s.gcr.io/kube-proxy-amd64:v1.10.3
[root@host01-2 hk]# docker load -i weave-kube
5bef08742407: Loading layer [==================================================>] 4.221 MB/4.221 MB
c3355c8b5c3e: Loading layer [==================================================>] 19.03 MB/19.03 MB
a83fa3df4138: Loading layer [==================================================>] 29.55 MB/29.55 MB
020fdc01af85: Loading layer [==================================================>]  11.6 MB/11.6 MB
2ea881a632b7: Loading layer [==================================================>] 2.048 kB/2.048 kB
396aa46bcbea: Loading layer [==================================================>] 35.09 MB/35.09 MB
Loaded image: docker.io/weaveworks/weave-kube:2.3.0
[root@host01-2 hk]# docker load -i weave-npc
8dccfe2dec8c: Loading layer [==================================================>] 2.811 MB/2.811 MB
3249ff6df12f: Loading layer [==================================================>] 40.52 MB/40.52 MB
3dc458d34b22: Loading layer [==================================================>]  2.56 kB/2.56 kB
Loaded image: docker.io/weaveworks/weave-npc:2.3.0
[root@host01-2 hk]#
 
cs


*Master Node에서 실행 :

1
2
3
4
5
kubeadm init
 
mkdir -p $HOME/.kube
cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
chown $(id -u):$(id -g) $HOME/.kube/config
cs


실행방법 :

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
[root@host01-4 ~]# docker images
REPOSITORY                                 TAG                 IMAGE ID            CREATED             SIZE
k8s.gcr.io/kube-proxy-amd64                v1.10.3             4261d315109d        3 days ago          97.1 MB
k8s.gcr.io/kube-apiserver-amd64            v1.10.3             e03746fe22c3        3 days ago          225 MB
k8s.gcr.io/kube-controller-manager-amd64   v1.10.3             40c8d10b2d11        3 days ago          148 MB
k8s.gcr.io/kube-scheduler-amd64            v1.10.3             353b8f1d102e        3 days ago          50.4 MB
docker.io/weaveworks/weave-npc             2.3.0               21545eb3d6f9        6 weeks ago         47.2 MB
docker.io/weaveworks/weave-kube            2.3.0               f15514acce73        6 weeks ago         96.8 MB
k8s.gcr.io/etcd-amd64                      3.1.12              52920ad46f5b        2 months ago        193 MB
k8s.gcr.io/k8s-dns-dnsmasq-nanny-amd64     1.14.8              c2ce1ffb51ed        4 months ago        41 MB
k8s.gcr.io/k8s-dns-sidecar-amd64           1.14.8              6f7f2dc7fab5        4 months ago        42.2 MB
k8s.gcr.io/k8s-dns-kube-dns-amd64          1.14.8              80cc5ea4b547        4 months ago        50.5 MB
[root@host01-4 ~]# clear
[root@host01-4 ~]# kubeadm init
[init] Using Kubernetes version: v1.10.3
[init] Using Authorization modes: [Node RBAC]
[preflight] Running pre-flight checks.
        [WARNING FileExisting-crictl]: crictl not found in system path
Suggestion: go get github.com/kubernetes-incubator/cri-tools/cmd/crictl
[certificates] Generated ca certificate and key.
[certificates] Generated apiserver certificate and key.
[certificates] apiserver serving cert is signed for DNS names [host01-4.cloud.com kubernetes kubernetes.default kubernetes.default.svc kubernetes.default.svc.cluster.local] and IPs [10.96.0.1 10.10.12.14]
[certificates] Generated apiserver-kubelet-client certificate and key.
[certificates] Generated etcd/ca certificate and key.
[certificates] Generated etcd/server certificate and key.
[certificates] etcd/server serving cert is signed for DNS names [localhost] and IPs [127.0.0.1]
[certificates] Generated etcd/peer certificate and key.
[certificates] etcd/peer serving cert is signed for DNS names [host01-4.cloud.com] and IPs [10.10.12.14]
[certificates] Generated etcd/healthcheck-client certificate and key.
[certificates] Generated apiserver-etcd-client certificate and key.
[certificates] Generated sa key and public key.
[certificates] Generated front-proxy-ca certificate and key.
[certificates] Generated front-proxy-client certificate and key.
[certificates] Valid certificates and keys now exist in "/etc/kubernetes/pki"
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/admin.conf"
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/kubelet.conf"
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/controller-manager.conf"
[kubeconfig] Wrote KubeConfig file to disk: "/etc/kubernetes/scheduler.conf"
[controlplane] Wrote Static Pod manifest for component kube-apiserver to "/etc/kubernetes/manifests/kube-apiserver.yaml"
[controlplane] Wrote Static Pod manifest for component kube-controller-manager to "/etc/kubernetes/manifests/kube-controller-manager.yaml"
[controlplane] Wrote Static Pod manifest for component kube-scheduler to "/etc/kubernetes/manifests/kube-scheduler.yaml"
[etcd] Wrote Static Pod manifest for a local etcd instance to "/etc/kubernetes/manifests/etcd.yaml"
[init] Waiting for the kubelet to boot up the control plane as Static Pods from directory "/etc/kubernetes/manifests".
[init] This might take a minute or longer if the control plane images have to be pulled.
[apiclient] All control plane components are healthy after 25.003852 seconds
[uploadconfig] Storing the configuration used in ConfigMap "kubeadm-config" in the "kube-system" Namespace
[markmaster] Will mark node host01-4.cloud.com as master by adding a label and a taint
[markmaster] Master host01-4.cloud.com tainted and labelled with key/value: node-role.kubernetes.io/master=""
[bootstraptoken] Using token: hlz7wp.qjrgmsq2yn9f94wa
[bootstraptoken] Configured RBAC rules to allow Node Bootstrap tokens to post CSRs in order for nodes to get long term certificate credentials
[bootstraptoken] Configured RBAC rules to allow the csrapprover controller automatically approve CSRs from a Node Bootstrap Token
[bootstraptoken] Configured RBAC rules to allow certificate rotation for all node client certificates in the cluster
[bootstraptoken] Creating the "cluster-info" ConfigMap in the "kube-public" namespace
[addons] Applied essential addon: kube-dns
[addons] Applied essential addon: kube-proxy
 
Your Kubernetes master has initialized successfully!
 
To start using your cluster, you need to run the following as a regular user:
 
  mkdir -p $HOME/.kube
  sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
  sudo chown $(id -u):$(id -g) $HOME/.kube/config
 
You should now deploy a pod network to the cluster.
Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:
  https://kubernetes.io/docs/concepts/cluster-administration/addons/
 
You can now join any number of machines by running the following on each node
as root:
 
  kubeadm join 10.10.12.14:6443 --token hlz7wp.qjrgmsq2yn9f94wa --discovery-token-ca-cert-hash sha256:43e61417b20ede5ca530fe0638990bc1a805b5f2a9e25b5aa2f40023b392fb50
 
[root@host01-4 ~]#
[root@host01-4 ~]# mkdir -p $HOME/.kube
[root@host01-4 ~]# cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
[root@host01-4 ~]# chown $(id -u):$(id -g) $HOME/.kube/config
[root@host01-4 ~]#
 
cs


위 실행결과에서 토큰을 복사해놓는다 :

1
2
3
  kubeadm join 10.10.12.14:6443 --token hlz7wp.qjrgmsq2yn9f94wa --discovery-token-ca-cert-hash sha256:43e61417b20ede5ca530fe0638990bc1a805b5f2a9e25b5aa2f40023b392fb50
 
 
cs


*Pod Network 설치 (마스터)

1
2
3
4
5
6
7
sysctl net.bridge.bridge-nf-call-iptables=1
 
export kubever=$(kubectl version | base64 | tr -d '\n')
 
kubectl apply -f "https://cloud.weave.works/k8s/net?k8s-version=$kubever"
 
kubectl get pods --all-namespaces 
cs


실행결과 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
[root@host01-4 ~]# sysctl net.bridge.bridge-nf-call-iptables=1
net.bridge.bridge-nf-call-iptables = 1
[root@host01-4 ~]#
[root@host01-4 ~]# export kubever=$(kubectl version | base64 | tr -d '\n')
[root@host01-4 ~]#
[root@host01-4 ~]# kubectl apply -f "https://cloud.weave.works/k8s/net?k8s-version=$kubever"
serviceaccount "weave-net" created
clusterrole.rbac.authorization.k8s.io "weave-net" created
clusterrolebinding.rbac.authorization.k8s.io "weave-net" created
role.rbac.authorization.k8s.io "weave-net" created
rolebinding.rbac.authorization.k8s.io "weave-net" created
daemonset.extensions "weave-net" created
[root@host01-4 ~]# kubectl get pods --all-namespaces
NAMESPACE     NAME                                         READY     STATUS    RESTARTS   AGE
kube-system   etcd-host01-4.cloud.com                      1/1       Running   0          4m
kube-system   kube-apiserver-host01-4.cloud.com            1/1       Running   0          4m
kube-system   kube-controller-manager-host01-4.cloud.com   1/1       Running   0          5m
kube-system   kube-dns-86f4d74b45-t9df2                    0/3       Pending   0          5m
kube-system   kube-proxy-fs9d8                             1/1       Running   0          5m
kube-system   kube-scheduler-host01-4.cloud.com            1/1       Running   0          4m
kube-system   weave-net-zr5qr                              2/2       Running   0          11s
[root@host01-4 ~]#
 
 
cs


*이전에 Master Node에서 복사한 토큰을 2개 Worker Node에 각각

입력한다 : 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
[root@host01-3 ~]# kubeadm join 10.10.12.14:6443 --token hlz7wp.qjrgmsq2yn9f94wa --discovery-token-ca-cert-hash sha256:43e61417b20ede5ca530fe0638990bc1a805b5f2a9e25b5aa2f40023b392fb50
[preflight] Running pre-flight checks.
        [WARNING FileExisting-crictl]: crictl not found in system path
Suggestion: go get github.com/kubernetes-incubator/cri-tools/cmd/crictl
[discovery] Trying to connect to API Server "10.10.12.14:6443"
[discovery] Created cluster-info discovery client, requesting info from "https://10.10.12.14:6443"
[discovery] Requesting info from "https://10.10.12.14:6443" again to validate TLS against the pinned public key
[discovery] Cluster info signature and contents are valid and TLS certificate validates against pinned roots, will use API Server "10.10.12.14:6443"
[discovery] Successfully established connection with API Server "10.10.12.14:6443"
 
This node has joined the cluster:
* Certificate signing request was sent to master and a response
  was received.
* The Kubelet was informed of the new secure connection details.
 
Run 'kubectl get nodes' on the master to see this node join the cluster.
[root@host01-3 ~]#
 
cs


*Master Node에서 멤버 Join이 이루어졌는지 확인한다(Status에 Ready가 떠야 한다.)

1
2
3
4
5
6
7
8
 
[root@host01-4 ~]# kubectl get no
NAME                 STATUS     ROLES     AGE       VERSION
host01-2.cloud.com   NotReady   <none>    6s        v1.10.3
host01-3.cloud.com   NotReady   <none>    11s       v1.10.3
host01-4.cloud.com   Ready      master    9m        v1.10.3
[root@host01-4 ~]# ^C
 
cs
1
2
3
4
5
6
7
[root@host01-4 ~]# kubectl get no
NAME                 STATUS    ROLES     AGE       VERSION
host01-2.cloud.com   Ready     <none>    2m        v1.10.3
host01-3.cloud.com   Ready     <none>    2m        v1.10.3
host01-4.cloud.com   Ready     master    11m       v1.10.3
[root@host01-4 ~]#
 
cs






*Tab 자동 입력


AWS, Azure에서는 kubectl 명령어가 그냥 'k'임(alias로 등록되어 있음)


*Alias 등록 : 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
[root@host01-4 ~]# alias k=kubectl
[root@host01-4 ~]# source <(kubectl completion bash | sed  s/kubectl/k/g)
[root@host01-4 ~]# k get
You must specify the type of resource to get. Valid resource types include:
 
  * all
  * certificatesigningrequests (aka 'csr')
  * clusterrolebindings
  * clusterroles
  * componentstatuses (aka 'cs')
  * configmaps (aka 'cm')
  * controllerrevisions
  * cronjobs
  * customresourcedefinition (aka 'crd')
  * daemonsets (aka 'ds')
  * deployments (aka 'deploy')
  * endpoints (aka 'ep')
  * events (aka 'ev')
  * horizontalpodautoscalers (aka 'hpa')
  * ingresses (aka 'ing')
  * jobs
  * limitranges (aka 'limits')
  * namespaces (aka 'ns')
  * networkpolicies (aka 'netpol')
  * nodes (aka 'no')
  * persistentvolumeclaims (aka 'pvc')
  * persistentvolumes (aka 'pv')
  * poddisruptionbudgets (aka 'pdb')
  * podpreset
  * pods (aka 'po')
  * podsecuritypolicies (aka 'psp')
  * podtemplates
  * replicasets (aka 'rs')
  * replicationcontrollers (aka 'rc')
  * resourcequotas (aka 'quota')
  * rolebindings
  * roles
  * secrets
  * serviceaccounts (aka 'sa')
  * services (aka 'svc')
  * statefulsets (aka 'sts')
  * storageclasses (aka 'sc')error: Required resource not specified.
Use "kubectl explain <resource>" for a detailed description of that resource (e.g. kubectl explain pods).
See 'kubectl get -h' for help and examples.
[root@host01-4 ~]# echo "alias l=kubectl" >> ~/.bashrc
[root@host01-4 ~]# echo "source <(kubectl completion bash | sed s/kubectl/k/g)" >> ~/.bashrc
 
cs





*Pod 개념

MS 아키텍처의 기초는 하나의 컨테이너에는 하나의 어플리케이션만 집어넣자!

- 네트워크 공유 : Pod로 구성하면 Container끼리 네트워크를 공유할 수 있게된다(네트워크 Name Space 공유 - localhost로)

- 스토리지 공유 : 볼륨을 pod에 마운트 하여 Storage까지 공유한다!


- 1개 pod 내에서 콘테이너들은 동일 호스트로만 구성이 된다. 똑같은 포트의 서비스를 공유할 수 없다.

(pod를 잘못구성하는 예 : DB Container와 Web Container를 동일 pod 내에 두었을 때 Scale Out 시 이슈가 있음,

따라서 web pod, WAS pod, DB pod 식으로 나눠야 함)

pod가 kubenetes의 최소 단위로 구성해야 함!


각 Work Node의 Pod 들은 Master Node의 replication controller(RC)에서 제어된다.


- 참조/관리 형태

참조를 하는 주체는 selector (key/value 형태)

참조를 받은 대상은 Label (key/value 형태)  = Pod(Node Selector)

Pod/Label의 참조를 받는 대상은 Label (key/value 형태) 

라고 부른다.

- Kubenetes의 Namespace란 : 


1
2
3
4
5
6
7
8
9
10
11
12
13
14
[root@host01-4 hk]# k get po -n kube-system -o  wide
NAME                                         READY     STATUS    RESTARTS   AGE       IP            NODE
etcd-host01-4.cloud.com                      1/1       Running   0          58m       10.10.12.14   host01-4.cloud.com
kube-apiserver-host01-4.cloud.com            1/1       Running   0          58m       10.10.12.14   host01-4.cloud.com
kube-controller-manager-host01-4.cloud.com   1/1       Running   0          59m       10.10.12.14   host01-4.cloud.com
kube-dns-86f4d74b45-t9df2                    3/3       Running   0          1h        10.32.0.2     host01-4.cloud.com
kube-proxy-fs9d8                             1/1       Running   0          1h        10.10.12.14   host01-4.cloud.com
kube-proxy-r5bzj                             1/1       Running   0          51m       10.10.12.13   host01-3.cloud.com
kube-proxy-tvwnv                             1/1       Running   0          51m       10.10.12.12   host01-2.cloud.com
kube-scheduler-host01-4.cloud.com            1/1       Running   0          59m       10.10.12.14   host01-4.cloud.com
weave-net-hf9d5                              2/2       Running   1          51m       10.10.12.12   host01-2.cloud.com
weave-net-p5drv                              2/2       Running   1          51m       10.10.12.13   host01-3.cloud.com
weave-net-zr5qr                              2/2       Running   0          54m       10.10.12.14   host01-4.cloud.com
 
cs

*Cluster PAI 확인 : 

1
2
3
4
5
[root@host01-4 hk]# k cluster-info
Kubernetes master is running at https://10.10.12.14:6443
KubeDNS is running at https://10.10.12.14:6443/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
 
To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.
cs


*Pod 가 어디있는지 상관없이 연결이 된다

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
[root@host01-4 hk]# ls
dns-kube-dns  etcd-amd64  kubeadm.host08-1.root.log.INFO.20180525-060140.2620  kube-apiserver   kube-proxy      weave-kube
dns-sidecar   k8s-dns     kubeadm.INFO                                         kube-controller  kube-scheduler  weave-npc
[root@host01-4 hk]# k run --image=reg.cloud.com/nginx --port=80 --generator=run/v1
error: NAME is required for run
See 'kubectl run -h' for help and examples.
[root@host01-4 hk]# k run --image=reg.cloud.com/nginx --port=80 --generator=run/v1
error: NAME is required for run
See 'kubectl run -h' for help and examples.
[root@host01-4 hk]# k run --image=reg.cloud.com/nginx nginx-app --port=80 --generator=run/v1
replicationcontroller "nginx-app" created
[root@host01-4 hk]# k get rc
NAME        DESIRED   CURRENT   READY     AGE
nginx-app   1         1         0         8s
[root@host01-4 hk]# k get po
NAME              READY     STATUS    RESTARTS   AGE
nginx-app-gb6ch   1/1       Running   0          15s
[root@host01-4 hk]# get po -o wide
-bash: get: command not found
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
nginx-app-gb6ch   1/1       Running   0          32s       10.36.0.1   host01-3.cloud.com
[root@host01-4 hk]# k logs
error: expected 'logs (POD | TYPE/NAME) [CONTAINER_NAME]'.
POD or TYPE/NAME is a required argument for the logs command
See 'kubectl logs -h' for help and examples.
[root@host01-4 hk]# k logs nginx-app-df618
Error from server (NotFound): pods "nginx-app-df618" not found
[root@host01-4 hk]# k logs nginx-app-gb6ch
[root@host01-4 hk]# k exec -it
error: expected 'exec POD_NAME COMMAND [ARG1] [ARG2] ... [ARGN]'.
POD_NAME and COMMAND are required arguments for the exec command
See 'kubectl exec -h' for help and examples.
[root@host01-4 hk]# k exec -it nginx-app-gb6ch bash
root@nginx-app-gb6ch:/#
 
cs

-podname이 container의 hostname이 된다.



*pod name으로 삭제 해도 다시 살아난다. 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
nginx-app-gb6ch   1/1       Running   0          5m        10.36.0.1   host01-3.cloud.com
[root@host01-4 hk]# delete po nginx-app-gb6ch
-bash: delete: command not found
[root@host01-4 hk]# k delete po nginx-app-gb6ch
pod "nginx-app-gb6ch" deleted
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS              RESTARTS   AGE       IP        NODE
nginx-app-gnpsd   0/1       ContainerCreating   0          6s        <none>    host01-2.cloud.com
[root@host01-4 hk]# k get rc
NAME        DESIRED   CURRENT   READY     AGE
nginx-app   1         1         1         5m
[root@host01-4 hk]#
 
cs


*Scale in Scale out


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
[root@host01-4 hk]# k describe rc nginx-app
dns-kube-dns                                         kubeadm.host08-1.root.log.INFO.20180525-060140.2620  kube-proxy
dns-sidecar                                          kubeadm.INFO                                         kube-scheduler
etcd-amd64                                           kube-apiserver                                       weave-kube
k8s-dns                                              kube-controller                                      weave-npc
[root@host01-4 hk]# k describe rc nginx-app
Name:         nginx-app
Namespace:    default
Selector:     run=nginx-app
Labels:       run=nginx-app
Annotations:  <none>
Replicas:     1 current / 1 desired
Pods Status:  1 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:  run=nginx-app
  Containers:
   nginx-app:
    Image:        reg.cloud.com/nginx
    Port:         80/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:       <none>
  Volumes:        <none>
Events:
  Type    Reason            Age   From                    Message
  ----    ------            ----  ----                    -------
  Normal  SuccessfulCreate  6m    replication-controller  Created pod: nginx-app-gb6ch
  Normal  SuccessfulCreate  1m    replication-controller  Created pod: nginx-app-gnpsd
[root@host01-4 hk]# k get po
NAME              READY     STATUS    RESTARTS   AGE
nginx-app-gnpsd   1/1       Running   0          1m
[root@host01-4 hk]# k describe po nginx-app-gnpsd
Name:           nginx-app-gnpsd
Namespace:      default
Node:           host01-2.cloud.com/10.10.12.12
Start Time:     Fri, 25 May 2018 11:58:20 +0900
Labels:         run=nginx-app
Annotations:    <none>
Status:         Running
IP:             10.44.0.1
Controlled By:  ReplicationController/nginx-app
Containers:
  nginx-app:
    Container ID:   docker://6d0e9cb190b31334dee5dba4877ace52d8afd5a9956d7c50eae35d3107722a58
    Image:          reg.cloud.com/nginx
    Image ID:       docker-pullable://reg.cloud.com/nginx@sha256:a4fb15454c43237dbc6592c4f8e0b50160ceb03e852a10c9895cf2a6d16c7fe2
    Port:           80/TCP
    Host Port:      0/TCP
    State:          Running
      Started:      Fri, 25 May 2018 11:58:29 +0900
    Ready:          True
    Restart Count:  0
    Environment:    <none>
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-85hdm (ro)
Conditions:
  Type           Status
  Initialized    True
  Ready          True
  PodScheduled   True
Volumes:
  default-token-85hdm:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-85hdm
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type    Reason                 Age   From                         Message
  ----    ------                 ----  ----                         -------
  Normal  SuccessfulMountVolume  1m    kubelet, host01-2.cloud.com  MountVolume.SetUp succeeded for volume "default-token-85hdm"
  Normal  Scheduled              1m    default-scheduler            Successfully assigned nginx-app-gnpsd to host01-2.cloud.com
  Normal  Pulling                1m    kubelet, host01-2.cloud.com  pulling image "reg.cloud.com/nginx"
  Normal  Pulled                 1m    kubelet, host01-2.cloud.com  Successfully pulled image "reg.cloud.com/nginx"
  Normal  Created                1m    kubelet, host01-2.cloud.com  Created container
  Normal  Started                1m    kubelet, host01-2.cloud.com  Started container
[root@host01-4 hk]# k get rc
NAME        DESIRED   CURRENT   READY     AGE
nginx-app   1         1         1         8m
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
nginx-app-gnpsd   1/1       Running   0          3m        10.44.0.1   host01-2.cloud.com
 
[root@host01-4 hk]# k scale rc nginx-app --replicas=3
replicationcontroller "nginx-app" scaled
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
nginx-app-gnpsd   1/1       Running   0          4m        10.44.0.1   host01-2.cloud.com
nginx-app-jfmkd   1/1       Running   0          10s       10.44.0.2   host01-2.cloud.com
nginx-app-ww6sn   1/1       Running   0          10s       10.36.0.1   host01-3.cloud.com
 
[root@host01-4 hk]# k scale rc nginx-app --replicas=0
replicationcontroller "nginx-app" scaled
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS        RESTARTS   AGE       IP          NODE
nginx-app-gnpsd   0/1       Terminating   0          5m        10.44.0.1   host01-2.cloud.com
nginx-app-jfmkd   0/1       Terminating   0          34s       10.44.0.2   host01-2.cloud.com
[root@host01-4 hk]# k get po -o wide
No resources found.
[root@host01-4 hk]# rc 0
-bash: rc: command not found
[root@host01-4 hk]# k scale rc nginx-app --replicas=1
replicationcontroller "nginx-app" scaled
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
nginx-app-7qpbv   1/1       Running   0          4s        10.36.0.1   host01-3.cloud.com
[root@host01-4 hk]#
 
cs



*Yaml으로 pod 조회

1
2
[root@host01-4 hk]# k get po nginx-app-7qpbv -o yaml
 
cs


*Yaml 파일로 추출 후 pod 생성

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
apiVersion: v1
kind: Pod
metadata:
  labels:
    type: web
  name: nginx-hk-app
spec:
  containers:
  - image: reg.cloud.com/nginx
    name: nginx-app
    ports:                            #expose와 동일
    - containerPort: 80
      protocol: TCP
 
 
 
[root@host01-4 hk]# k get po nginx-app-7qpbv -o yaml > temp.yaml
[root@host01-4 hk]# vi temp.yaml
[root@host01-4 hk]# k create -f temp.yaml
pod "nginx-hk-app" created
[root@host01-4 hk]# k get po -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP          NODE
nginx-app-7qpbv   1/1       Running   0          6m        10.36.0.1   host01-3.cloud.com
nginx-hk-app      1/1       Running   0          13s       10.44.0.1   host01-2.cloud.com
[root@host01-4 hk]#
 
cs



*Yaml 파일로 rc 생성 후 배포 :


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
[root@host01-4 hk]# k get rc nginx-app -o yaml > hk.yaml
[root@host01-4 hk]# vi hk.yaml
apiVersion: v1
kind: ReplicationController
metadata:
  labels:
    run: nginx-app
  name: nginx-app
spec:
  replicas: 1
  selector:
    type: test
  template:
    metadata:
      labels:
        type: test
    spec:
      containers:
      - image: reg.cloud.com/nginx
        name: nginx-app
        ports:
        - containerPort: 80
          protocol: TCP
 
 
[root@host01-4 hk]# k create -f hk.yaml
replicationcontroller "nginx-app2" created
[root@host01-4 hk]# k get rc
NAME         DESIRED   CURRENT   READY     AGE
nginx-app    1         1         1         24m
nginx-app2   1         1         1         9s
[root@host01-4 hk]# k get po -o wide
NAME               READY     STATUS    RESTARTS   AGE       IP          NODE
nginx-app-7qpbv    1/1       Running   0          14m       10.36.0.1   host01-3.cloud.com
nginx-app2-tgqqf   1/1       Running   0          16s       10.36.0.2   host01-3.cloud.com
nginx-hk-app       1/1       Running   0          7m        10.44.0.1   host01-2.cloud.com
[root@host01-4 hk]#
 
cs


저작자표시

'Docker(도커) Kubernetes' 카테고리의 다른 글

Docker Kubernetes, Readiness Probe 생성 방법  (0) 2018.05.25
Pod 개념, 멀티 컨테이너 생성, kubectl 명령어 사용법  (0) 2018.05.25
도커 관련 설정 파일 모음  (2) 2018.05.24
docker swarm 사용법  (0) 2018.05.24
Docker cp 명령어, Docker Compose  (0) 2018.05.24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
yum -y install bash-completion
 
curl -sSL https://get.docker.com/ | sh
 
 
systemctl start docker
 
===cAdvisor=================================
docker run \
   --detach=true \
   --volume=/:/rootfs:ro \
   --volume=/var/run:/var/run:rw \
   --volume=/sys:/sys:ro \
   --volume=/var/lib/docker/:/var/lib/docker:ro \
   --publish=8080:8080 \
   --privileged=true \
   --name=cadvisor \
google/cadvisor:latest
 
docker service create \
   --mode global \
   --mount type=bind,source=/,destination=/rootfs,ro=1 \
   --mount type=bind,source=/var/run,destination=/var/run \
   --mount type=bind,source=/sys,destination=/sys,ro=1 \
   --mount type=bind,source=/var/lib/docker,destination=/var/lib/docker,ro=1 \
   --publish mode=host,target=8080,published=8080 \
   --name=cadvisor \
   google/cadvisor:latest
 
===Compose=========
curl -L https://github.com/docker/compose/releases/download/1.21.0/docker-compose-`uname -s`-`uname -m` -o /usr/local/bin/docker-compose
 
chmod +x /usr/local/bin/docker-compose
 
cs



1
2
3
4
5
6
{
"insecure-registries":["10.10.12.0/24"]
}
 
 
/var/dockers/daemon.json
cs



1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
version: '3.3'
 
services:
   db:
     image: reg.cloud.com/mysql:5.7
     volumes:
       - dbdata:/var/lib/mysql
     restart: always
     environment:
       MYSQL_ROOT_PASSWORD: somewordpress
       MYSQL_DATABASE: wordpress
       MYSQL_USER: wordpress
       MYSQL_PASSWORD: wordpress
 
   wordpress:
     depends_on:
       - db
     image: reg.cloud.com/wordpress:latest
     ports:
       - "8000:80"
     restart: always
     environment:
       WORDPRESS_DB_HOST: db:3306
       WORDPRESS_DB_USER: wordpress
       WORDPRESS_DB_PASSWORD: wordpress
volumes:
    dbdata:
cs





저작자표시

'Docker(도커) Kubernetes' 카테고리의 다른 글

Pod 개념, 멀티 컨테이너 생성, kubectl 명령어 사용법  (0) 2018.05.25
Kubernetes 설치 및 사용법, Yaml으로 pod 및 rc 생성  (0) 2018.05.25
docker swarm 사용법  (0) 2018.05.24
Docker cp 명령어, Docker Compose  (0) 2018.05.24
docker network 네트워크 연결 및 구성  (0) 2018.05.24

+ Recent posts

티스토리툴바