kubernetes
20210908 (수) pod scheduler
gusalstm
2021. 9. 8. 15:45
반응형
9:40
Pod Scheduler
| node name | manifests에서 수동으로 pods를 배치할 노드를 지정 - 스케쥴러의 관여 없음 scale 명령어로 scale out을 실행해도 같은 노드에서 배정됨 (replicaset에 설정된 node name 으로 인하여) |
| node selector | 노드의 레이블 지우기 (node 라는 label 항목을 마이너스(-)하는 명령) $ kubectl label node kube-node2 node- |
| affinity | 파드가 특정 노드에 배치되는 것을 선호하는 노드를 지정 node affinity : pod를 특정 node에 배치하는것을 선호 pod affinity : pod 간에 같은 노드에 배치하는 것을 선호 pod anti-affinity : pod 간에 다른 노드에 배치하는 것을 선호 |
| - node affinity | node affinity : node selector와 유사하게 pod를 특정 node에 배치할 수 있음 node selector는 해당 label이 지정된 node에만 배치되지만 node affinity는 특정 node에 배치하는 것을 강제하지 않을 수 있음 nodeAffinity 유형과 연관된 nodeSelectorTerms 를 지정하면, nodeSelectorTerms 중 하나라도 만족시키는 노드에 파드가 스케줄된다. nodeSelectorTerms 와 연관된 여러 matchExpressions 를 지정하면, 파드는 matchExpressions 를 모두 만족하는 노드에만 스케줄된다. |
| - pod affinity | pod affinity |
| - pod anti-affinity | pod anti-affinity scale in scale out시에도 지켜짐. |
| taints & tolerations | taints : 특정 노드에 더이상 pod가 추가되지 않도록 pod를 스케쥴링 하지 않는것. → 별도의 tolerations이 없는 경우 pod가 추가 생성되지 않음 → 정의형태 : "KEY[=VALUE]:EFFECT" taints effect의 종류 : - NoSchedule : tolerations이 정의되어있지 않은경우 pod를 스케쥴링 하지 않음 - PreferNoSchedule : NoSchedule과 동일하나 리소스 부족 등의 예외상황에 배치할 수 있음 (기존 pod에 적용되지 않음) - NoExecute : tolerations이 정의되어있지 않은경우 pod를 스케쥴링 할 수 없음 (기존 pod도 설정적용을 받음) tolerations : taints가 적용된 노드에 pod를 배치할 수 있도록 하는것 vagrant@kube-control1:~/work/20210908$ kubectl describe nodes |grep Taints Taints: node-role.kubernetes.io/master:NoSchedule → control plane은 적용되어있음 Taints: <none> Taints: <none> Taints: <none> |
| cordon & drain |
### nodename : kube-node1에만 생성 ###
vagrant@kube-control1:~/work/20210908$ cat myapp-rs-nodename.yaml
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: myapp-rs-nodename
spec:
replicas: 2
selector:
matchLabels:
app: myapp-rs-nodename
template:
metadata:
labels:
app: myapp-rs-nodename
spec:
nodeName: kube-node1
containers:
- name: myapp
image: devops2341/go-myweb:latest
### node selector : gpu:highend라는 레이블을 가진 node에만 생성 ###
vagrant@kube-control1:~/work/20210908$ cat myapp-rs-nodeselector.yaml
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: myapp-rs-nodeselector
spec:
replicas: 2
selector:
matchLabels:
app: myapp-rs-nodeselector
template:
metadata:
labels:
app: myapp-rs-nodeselector
spec:
nodeSelector:
gpu: highend
containers:
- name: myapp
image: devops2341/go-myweb:latest
vagrant@kube-control1:~/work/20210908$ kubectl get nodes --show-labels
NAME STATUS ROLES AGE VERSION LABELS
kube-control1 Ready master 15d v1.19.11 kubernetes.io
kube-node1 Ready <none> 15d v1.19.11 gpu=highend #다른 레이블 생략함
kube-node2 Ready <none> 15d v1.19.11 gpu=midrange#다른 레이블 생략함
kube-node3 Ready <none> 15d v1.19.11 gpu=lowends #다른 레이블 생략함
vagrant@kube-control1:~/work/20210908$ kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE NODE
myapp-rs-nodeselector-mfz5f 1/1 Running 0 10s kube-node1
myapp-rs-nodeselector-rqvrn 1/1 Running 0 10s kube-node1
# node1 : gpu=highend 에만 생성 #
### node affinity : 기본 문법 ###
spec:
affinity:
nodeAffinity:
requireDuringSchedulingIgnoredDuringExecution: #반드시 조건 만족 (필수조건)
nodeSelectorTerms:
- matchExpressions:
- key: gpu-model # label name
operator: in # 조건연산자
values:
- '3080' # label value
- '2080' # label value
perferredDuringSchedulingIgnoredDuringExecution: #조건 선호 (예외 가능)
- weight: 10 # 가중치 ( 1 - 100 )
perferences:
matchLabels:
- key: gpu-model
operator: in
values:
- titan
### pod affinity : 기본 문법 ###
spec:
affinity:
podAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: KEY
operator: In
values:
- 'val1'
- 'val2'
topologyKey: TOPOLOGY_KEY
# topologykey: pod affinity, pod anti-affinity 파드를 배치/분리하는 기준
# 보통 노드의 이름을 참조하여 파드 배치/분리를 결정
### pod anti-affinity : 기본 문법 ###
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: KEY
operator: In
values:
- 'val1'
topologyKey: TOPOLOGY_KEY
### no notolerations.###
vagrant@kube-control1:~/work/20210908$ cat myapp-rs-podaff-cache.yaml
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: myapp-rs-aff-cache
spec:
replicas: 2
selector:
matchLabels:
app: myapp-rs-aff-cache
tier: cache
template:
metadata:
labels:
app: myapp-rs-aff-cache
tier: cache
spec:
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: tier
operator: In
values:
- cache
topologyKey: "kubernetes.io/hostname"
containers:
- name: myapp
image: devops2341/go-myweb:latest
vagrant@kube-control1:~/work/20210908$ cat myapp-rs-podaff-front.yaml
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: myapp-rs-aff-front
spec:
replicas: 2
selector:
matchLabels:
app: myapp-rs-aff-front
tier: frontend
template:
metadata:
labels:
app: myapp-rs-aff-front
tier: frontend
spec:
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: tier
operator: In
values:
- frontend
topologyKey: "kubernetes.io/hostname"
podAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: tier
operator: In
values:
- cache
topologyKey: "kubernetes.io/hostname"
containers:
- name: myapp
image: devops2341/go-myweb:latest
vagrant@kube-control1:~/work/20210908$
vagrant@kube-control1:~/work/20210908$ cat myapp-rs-notoleration.yaml
apiVersion: apps/v1
kind: ReplicaSet
metadata:
name: myapp-rs-notol
spec:
replicas: 1
selector:
matchLabels:
app: myapp-rs-notol
tier: backend
template:
metadata:
labels:
app: myapp-rs-notol
tier: backend
spec:
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: tier
operator: In
values:
- cache
topologyKey: "kubernetes.io/hostname"
containers:
- name: myapp
image: devops2341/go-myweb:latest
vagrant@kube-control1:~/work/20210908$
vagrant@kube-control1:~/work/20210908$ kubectl scale replicaset myapp-rs-aff-cache --replicas 3
replicaset.apps/myapp-rs-aff-cache scaled
vagrant@kube-control1:~/work/20210908$ kubectl scale replicaset myapp-rs-aff-front --replicas 3
replicaset.apps/myapp-rs-aff-front scaled
vagrant@kube-control1:~/work/20210908$ kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
myapp-rs-aff-cache-2rr5x 1/1 Running 0 23s 192.168.233.253 kube-node2 <none> <none>
myapp-rs-aff-cache-5g8kv 1/1 Running 0 3m36s 192.168.9.112 kube-node1 <none> <none>
myapp-rs-aff-cache-79n69 1/1 Running 0 3m36s 192.168.119.135 kube-node3 <none> <none>
myapp-rs-aff-front-dcpfl 1/1 Running 0 119s 192.168.119.138 kube-node3 <none> <none>
myapp-rs-aff-front-qts4s 1/1 Running 0 119s 192.168.9.116 kube-node1 <none> <none>
myapp-rs-aff-front-zq65m 1/1 Running 0 6s 192.168.233.251 kube-node2 <none> <none>
myapp-rs-nodeselector-mfz5f 1/1 Running 0 162m 192.168.9.101 kube-node1 <none> <none>
myapp-rs-nodeselector-rqvrn 1/1 Running 0 162m 192.168.9.98 kube-node1 <none> <none>
vagrant@kube-control1:~/work/20210908$ kubectl scale replicaset myapp-rs-aff-cache --replicas 2
replicaset.apps/myapp-rs-aff-cache scaled
vagrant@kube-control1:~/work/20210908$ kubectl scale replicaset myapp-rs-aff-front --replicas 2
replicaset.apps/myapp-rs-aff-front scaled
vagrant@kube-control1:~/work/20210908$ kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
myapp-rs-aff-cache-5g8kv 1/1 Running 0 5m33s 192.168.9.112 kube-node1 <none> <none>
myapp-rs-aff-cache-79n69 1/1 Running 0 5m33s 192.168.119.135 kube-node3 <none> <none>
myapp-rs-aff-front-dcpfl 1/1 Running 0 3m56s 192.168.119.138 kube-node3 <none> <none>
myapp-rs-aff-front-qts4s 1/1 Running 0 3m56s 192.168.9.116 kube-node1 <none> <none>
myapp-rs-nodeselector-mfz5f 1/1 Running 0 164m 192.168.9.101 kube-node1 <none> <none>
myapp-rs-nodeselector-rqvrn 1/1 Running 0 164m 192.168.9.98 kube-node1 <none> <none>
vagrant@kube-control1:~/work/20210908$ kubectl get nodes --show-labels
NAME STATUS ROLES AGE VERSION LABELS
kube-control1 Ready master 15d v1.19.11 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=kube-control1,kubernetes.io/os=linux,node-role.kubernetes.io/master=
kube-node1 Ready <none> 15d v1.19.11 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,gpu-model=3080,gpu=highend,kubernetes.io/arch=amd64,kubernetes.io/hostname=kube-node1,kubernetes.io/os=linux
kube-node2 Ready <none> 15d v1.19.11 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,gpu-model=2080,gpu=midrange,kubernetes.io/arch=amd64,kubernetes.io/hostname=kube-node2,kubernetes.io/os=linux
kube-node3 Ready <none> 15d v1.19.11 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,gpu-model=1660,gpu=lowends,kubernetes.io/arch=amd64,kubernetes.io/hostname=kube-node3,kubernetes.io/os=linux
vagrant@kube-control1:~/work/20210908$ kubectl describe nodes kube-nodes |grep Taints
Error from server (NotFound): nodes "kube-nodes" not found
vagrant@kube-control1:~/work/20210908$ kubectl describe nodes kube-nodes |grep -iname Taints
grep: invalid max count
Error from server (NotFound): nodes "kube-nodes" not found
vagrant@kube-control1:~/work/20210908$ kubectl describe nodes |grep -iname Taints
grep: invalid max count
vagrant@kube-control1:~/work/20210908$ kubectl describe nodes |grep taints
vagrant@kube-control1:~/work/20210908$ kubectl describe nodes |grep Taints
Taints: node-role.kubernetes.io/master:NoSchedule
Taints: <none>
Taints: <none>
Taints: <none>
vagrant@kube-control1:~/work/20210908$ kubectl taint NODE NODE-NAME KEY=[VALUE]:EFFECT ^C
vagrant@kube-control1:~/work/20210908$
vagrant@kube-control1:~/work/20210908$ kubectl get node --show-labels
NAME STATUS ROLES AGE VERSION LABELS
kube-control1 Ready master 15d v1.19.11 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=kube-control1,kubernetes.io/os=linux,node-role.kubernetes.io/master=
kube-node1 Ready <none> 15d v1.19.11 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,gpu-model=3080,gpu=highend,kubernetes.io/arch=amd64,kubernetes.io/hostname=kube-node1,kubernetes.io/os=linux
kube-node2 Ready <none> 15d v1.19.11 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,gpu-model=2080,gpu=midrange,kubernetes.io/arch=amd64,kubernetes.io/hostname=kube-node2,kubernetes.io/os=linux
kube-node3 Ready <none> 15d v1.19.11 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,gpu-model=1660,gpu=lowends,kubernetes.io/arch=amd64,kubernetes.io/hostname=kube-node3,kubernetes.io/os=linux
vagrant@kube-control1:~/work/20210908$ kubectl taint node kube-node3 env=production:NoSchedule
node/kube-node3 tainted
vagrant@kube-control1:~/work/20210908$ kubectl describe node kube-node3 |grep taint
vagrant@kube-control1:~/work/20210908$ kubectl describe node kube-node3
Name: kube-node3
Roles: <none>
Labels: beta.kubernetes.io/arch=amd64
beta.kubernetes.io/os=linux
gpu=lowends
gpu-model=1660
kubernetes.io/arch=amd64
kubernetes.io/hostname=kube-node3
kubernetes.io/os=linux
Annotations: csi.volume.kubernetes.io/nodeid: {"rook-ceph.cephfs.csi.ceph.com":"kube-node3","rook-ceph.rbd.csi.ceph.com":"kube-node3"}
kubeadm.alpha.kubernetes.io/cri-socket: /var/run/dockershim.sock
node.alpha.kubernetes.io/ttl: 0
projectcalico.org/IPv4Address: 192.168.200.23/24
projectcalico.org/IPv4IPIPTunnelAddr: 192.168.119.128
volumes.kubernetes.io/controller-managed-attach-detach: true
CreationTimestamp: Mon, 23 Aug 2021 07:52:59 +0000
Taints: env=production:NoSchedule
Unschedulable: false
Lease:
HolderIdentity: kube-node3
AcquireTime: <unset>
RenewTime: Wed, 08 Sep 2021 05:33:55 +0000
Conditions:
Type Status LastHeartbeatTime LastTransitionTime Reason
Message
---- ------ ----------------- ------------------ ------
-------
NetworkUnavailable False Wed, 08 Sep 2021 00:40:38 +0000 Wed, 08 Sep 2021 00:40:38 +0000 CalicoIsUp
Calico is running on this node
MemoryPressure False Wed, 08 Sep 2021 05:31:30 +0000 Wed, 08 Sep 2021 00:39:35 +0000 KubeletHasSufficientMemory kubelet has sufficient memory available
DiskPressure False Wed, 08 Sep 2021 05:31:30 +0000 Wed, 08 Sep 2021 00:39:35 +0000 KubeletHasNoDiskPressure kubelet has no disk pressure
PIDPressure False Wed, 08 Sep 2021 05:31:30 +0000 Wed, 08 Sep 2021 00:39:35 +0000 KubeletHasSufficientPID kubelet has sufficient PID available
Ready True Wed, 08 Sep 2021 05:31:30 +0000 Wed, 08 Sep 2021 00:39:36 +0000 KubeletReady
kubelet is posting ready status. AppArmor enabled
Addresses:
InternalIP: 192.168.200.23
Hostname: kube-node3
Capacity:
cpu: 2
ephemeral-storage: 40593612Ki
hugepages-2Mi: 0
memory: 3064356Ki
pods: 110
Allocatable:
cpu: 2
ephemeral-storage: 37411072758
hugepages-2Mi: 0
memory: 2961956Ki
pods: 110
System Info:
Machine ID: 6f1ad46bead544d1b5bcb11fd9fa3de9
System UUID: 3ee51e77-8110-6a4e-9a2d-52e44968fbd9
Boot ID: 7940543d-9dd5-4562-aff3-375558009495
Kernel Version: 5.4.0-81-generic
OS Image: Ubuntu 20.04.2 LTS
Operating System: linux
Architecture: amd64
Container Runtime Version: docker://20.10.8
Kubelet Version: v1.19.11
Kube-Proxy Version: v1.19.11
PodCIDR: 192.168.3.0/24
PodCIDRs: 192.168.3.0/24
Non-terminated Pods: (14 in total)
Namespace Name CPU Requests CPU Limits Memory Requests Memory Limits AGE
--------- ---- ------------ ---------- --------------- ------------- ---
default myapp-rs-aff-cache-79n69 0 (0%) 0 (0%) 0 (0%) 0 (0%) 93m
default myapp-rs-aff-front-dcpfl 0 (0%) 0 (0%) 0 (0%) 0 (0%) 91m
kube-system calico-node-4hg7w 250m (12%) 0 (0%) 0 (0%) 0 (0%) 15d
kube-system kube-proxy-pcpsn 0 (0%) 0 (0%) 0 (0%) 0 (0%) 15d
kube-system metrics-server-766c9b8df-2wwfc 100m (5%) 0 (0%) 200Mi (6%) 0 (0%) 28h
metallb-system speaker-bvsnw 0 (0%) 0 (0%) 0 (0%) 0 (0%) 14d
kind: ReplicaSet
rook-ceph csi-cephfsplugin-8fnps 0 (0%) 0 (0%) 0 (0%) 0 (0%) 5d4h
rook-ceph csi-cephfsplugin-provisioner-78d66674d8-c2xrl 0 (0%) 0 (0%) 0 (0%) 0 (0%) 5d4h
rook-ceph csi-rbdplugin-provisioner-687cf777ff-q9srr 0 (0%) 0 (0%) 0 (0%) 0 (0%) 5d4h
rook-ceph csi-rbdplugin-t586m 0 (0%) 0 (0%) 0 (0%)
metadata:
0 (0%) 5d4h
rook-ceph rook-ceph-crashcollector-kube-node3-fd9f9f6bd-hqtr4 0 (0%) 0 (0%) 0 (0%) 0 (0%) 29h
rook-ceph rook-ceph-mon-b-7dcc58fcc6-998r4 0 (0%) 0 (0%) 0 (0%) 0 (0%) 5d4h
rook-ceph rook-ceph-operator-fd9ff6bf5-8gsxf 0 (0%) 0 (0%) 0 (0%)
0 (0%) 5d4h
apiVersion: apps/v1
rook-ceph rook-ceph-osd-1-69c579789-r28h9 0 (0%) 0 (0%) 0 (0%) 0 (0%) 5d3h
Allocated resources:
(Total limits may be over 100 percent, i.e., overcommitted.)
Resource Requests Limits
-------- -------- ------
cpu 350m (17%) 0 (0%)
memory 200Mi (6%) 0 (0%)
ephemeral-storage 0 (0%) 0 (0%)
hugepages-2Mi 0 (0%) 0 (0%)
Events: <none>
vagrant@kube-control1:~/work/20210908$ kubectl describe node kube-node3 |grep Taint
Taints: env=production:NoSchedule
vagrant@kube-control1:~/work/20210908$ vim myapp-rs-notoleration.yaml
[New] 30L, 652C written
vagrant@kube-control1:~/work/20210908$ vim myapp-rs-notoleration.yaml
30L, 657C written
vagrant@kube-control1:~/work/20210908$ kubectl create -f myapp-rs-notoleration.yaml
replicaset.apps/myapp-rs-notol created
vagrant@kube-control1:~/work/20210908$ kubectl get replicasets
NAME DESIRED CURRENT READY AGE
myapp-rs-aff-cache 2 2 2 106m
myapp-rs-aff-front 2 2 2 105m
myapp-rs-nodeselector 2 2 2 4h25m
myapp-rs-notol 1 1 1 20s
vagrant@kube-control1:~/work/20210908$ kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
myapp-rs-aff-cache-5g8kv 1/1 Running 0 107m 192.168.9.112 kube-node1 <none> <none>
myapp-rs-aff-cache-79n69 1/1 Running 0 107m 192.168.119.135 kube-node3 <none> <none>
myapp-rs-aff-front-dcpfl 1/1 Running 0 105m 192.168.119.138 kube-node3 <none> <none>
myapp-rs-aff-front-qts4s 1/1 Running 0 105m 192.168.9.116 kube-node1 <none> <none>
myapp-rs-nodeselector-mfz5f 1/1 Running 0 4h26m 192.168.9.101 kube-node1 <none> <none>
myapp-rs-nodeselector-rqvrn 1/1 Running 0 4h26m 192.168.9.98 kube-node1 <none> <none>
myapp-rs-notol-77kfw 1/1 Running 0 28s 192.168.233.250 kube-node2 <none> <none>
vagrant@kube-control1:~/work/20210908$ kubectl delete -f myapp-rs-notoleration.yaml
replicaset.apps "myapp-rs-notol" deleted
vagrant@kube-control1:~/work/20210908$ kubectl delete -f myapp-rs-
myapp-rs-nodeaff.yaml myapp-rs-nodeselector.yaml myapp-rs-podaff-cache.yaml
myapp-rs-nodename.yaml myapp-rs-notoleration.yaml myapp-rs-podaff-front.yaml
vagrant@kube-control1:~/work/20210908$ kubectl delete -f myapp-rs-podaff-cache.yaml
replicaset.apps "myapp-rs-aff-cache" deleted
vagrant@kube-control1:~/work/20210908$ kubectl delete -f myapp-rs-podaff-front.yaml
replicaset.apps "myapp-rs-aff-front" deleted
vagrant@kube-control1:~/work/20210908$ kubectl get all
NAME READY STATUS RESTARTS AGE
pod/myapp-rs-nodeselector-mfz5f 1/1 Running 0 4h29m
pod/myapp-rs-nodeselector-rqvrn 1/1 Running 0 4h29m
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 15d
NAME DESIRED CURRENT READY AGE
replicaset.apps/myapp-rs-nodeselector 2 2 2 4h29m
vagrant@kube-control1:~/work/20210908$ kubectl create -f myapp-rs-notoleration.yaml
replicaset.apps/myapp-rs-notol created
vagrant@kube-control1:~/work/20210908$ kubectl delete -f myapp-rs-notoleration.yaml
replicaset.apps "myapp-rs-notol" deleted
vagrant@kube-control1:~/work/20210908$ kubectl create -f myapp-rs-
myapp-rs-nodeaff.yaml myapp-rs-nodeselector.yaml myapp-rs-podaff-cache.yaml
myapp-rs-nodename.yaml myapp-rs-notoleration.yaml myapp-rs-podaff-front.yaml
vagrant@kube-control1:~/work/20210908$ kubectl create -f myapp-rs-podaff-cache.yaml
replicaset.apps/myapp-rs-aff-cache created
vagrant@kube-control1:~/work/20210908$ kubectl create -f myapp-rs-podaff-front.yaml
replicaset.apps/myapp-rs-aff-front created
vagrant@kube-control1:~/work/20210908$ kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
myapp-rs-aff-cache-p7f8p 1/1 Running 0 7s 192.168.9.121 kube-node1 <none>
<none>
myapp-rs-aff-cache-zmqdx 1/1 Running 0 7s 192.168.233.255 kube-node2 <none>
<none>
myapp-rs-aff-front-7hdzp 0/1 ContainerCreating 0 4s <none> kube-node2 <none>
<none>
myapp-rs-aff-front-rwqwf 0/1 ContainerCreating 0 4s <none> kube-node1 <none>
<none>
myapp-rs-nodeselector-mfz5f 1/1 Running 0 4h29m 192.168.9.101 kube-node1 <none>
<none>
myapp-rs-nodeselector-rqvrn 1/1 Running 0 4h29m 192.168.9.98 kube-node1 <none>
<none>
vagrant@kube-control1:~/work/20210908$ kubectl create -f myapp-rs-notoleration.yaml
replicaset.apps/myapp-rs-notol created
vagrant@kube-control1:~/work/20210908$ kubectl get replicasets
NAME DESIRED CURRENT READY AGE
myapp-rs-aff-cache 2 2 2 25s
myapp-rs-aff-front 2 2 2 22s
myapp-rs-nodeselector 2 2 2 4h30m
myapp-rs-notol 1 1 0 4s
vagrant@kube-control1:~/work/20210908$ kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
myapp-rs-aff-cache-p7f8p 1/1 Running 0 34s 192.168.9.121 kube-node1 <none> <none>
myapp-rs-aff-cache-zmqdx 1/1 Running 0 34s 192.168.233.255 kube-node2 <none> <none>
myapp-rs-aff-front-7hdzp 1/1 Running 0 31s 192.168.233.194 kube-node2 <none> <none>
myapp-rs-aff-front-rwqwf 1/1 Running 0 31s 192.168.9.109 kube-node1 <none> <none>
myapp-rs-nodeselector-mfz5f 1/1 Running 0 4h30m 192.168.9.101 kube-node1 <none> <none>
myapp-rs-nodeselector-rqvrn 1/1 Running 0 4h30m 192.168.9.98 kube-node1 <none> <none>
myapp-rs-notol-tl9vg 0/1 Pending 0 13s <none> <none> <none> <none>
vagrant@kube-control1:~/work/20210908$ kubectl describe pods myapp-rs-notol-tl9vg
Name: myapp-rs-notol-tl9vg
Namespace: default
Priority: 0
Node: <none>
Labels: app=myapp-rs-notol
tier=backend
Annotations: <none>
Status: Pending
IP:
IPs: <none>
Controlled By: ReplicaSet/myapp-rs-notol
Containers:
myapp:
Image: devops2341/go-myweb:latest
Port: <none>
Host Port: <none>
Environment: <none>
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from default-token-w26v7 (ro)
Conditions:
Type Status
PodScheduled False
Volumes:
default-token-w26v7:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-w26v7
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling 63s (x3 over 2m7s) default-scheduler 0/4 nodes are available: 1 node(s) had taint {env: production}, that the pod didn't tolerate, 1 node(s) had taint {node-role.kubernetes.io/master: }, that the pod didn't tolerate, 2 node(s) didn't match pod affinity/anti-affinity, 2 node(s) didn't match pod anti-affinity rules.
vagrant@kube-control1:~/work/20210908$cordon & uncordon 명령어 kubectl cordon NODE kubectl uncordon NODE vagrant@kube-control1:~/work/20210908$ kubectl cordon kube-node1 node/kube-node1 cordoned vagrant@kube-control1:~/work/20210908$ kubectl cordon kube-node2 node/kube-node2 cordoned vagrant@kube-control1:~/work/20210908$ kubectl get nodes NAME STATUS ROLES AGE VERSION kube-control1 Ready master 15d v1.19.11 kube-node1 Ready,SchedulingDisabled <none> 15d v1.19.11 kube-node2 Ready,SchedulingDisabled <none> 15d v1.19.11 kube-node3 Ready <none> 15d v1.19.11 vagrant@kube-control1:~/work/20210908$ kubectl drain kube-node3 --ignore-daemonsets=true --delete-local-data=true node/kube-node3 already cordoned |
728x90