Kubectl: различия между версиями

Материал из noname.com.ua
Перейти к навигацииПерейти к поиску
 
(не показано 12 промежуточных версий этого же участника)
Строка 2: Строка 2:
 
[[Категория:kubectl]]
  
[[Категория:kubectl]]
 
=Шпаргалка по <code>kubectl</code>=
  
=Шпаргалка по <code>kubectl</code>=
  +
много примеров тут: https://kubernetes.io/docs/reference/kubectl/quick-reference/
 
==Частные случаи==
  
==Частные случаи==
 
===<code>--show-managed-fields </code> - посмотреть какие были операции с объектом===
  
===<code>--show-managed-fields </code> - посмотреть какие были операции с объектом===
Строка 104: Строка 105:
 
===Список контейнеров в поде===
  
===Список контейнеров в поде===
 
Для каждого контейнера pod-name, container-image container-name<BR>
  
Для каждого контейнера pod-name, container-image container-name<BR>
−
Довольно коряво, конечно
 +
Довольно коряво, конечно, но можно использовать как основу для своих скриптов
 
<PRE>
  
<PRE>
   
Строка 111: Строка 112:
 
-n osh-system \
  
-n osh-system \
 
--selector=app.kubernetes.io/instance=openstack-operator \
  
--selector=app.kubernetes.io/instance=openstack-operator \
−
-o=jsonpath='{range .items[*]}{"\nStart output for POD: "}{.metadata.name}{"\nContaines:\n"}{range .spec.containers[*]}{.image}{"\t"}{.name}{"\n"}{end}{"End output for POD"}{.metadata.name}{\n"}{end}'
 +
-o=jsonpath='{range .items[*]}{"\nStart output for POD: "}{.metadata.name}{"\nContaines:\n"}{range .spec.containers[*]}{.image}{"\t"}{.name}{"\n"}{end}{"End output for POD: "}{.metadata.name}{"\n"}{end}'
   
 
</PRE>
  
</PRE>
 
Пример вывода:
  
Пример вывода:
 
<PRE>
  
<PRE>
−
openstack-controller-7896bc6d4d-l5mrz 127.0.0.1:44301/openstack/openstack-controller:0.14.18 osdpl
 +
Start output for POD: openstack-controller-7896bc6d4d-l5mrz
  +
Containes:
  +
127.0.0.1:44301/openstack/openstack-controller:0.14.18 osdpl
 
127.0.0.1:44301/openstack/openstack-controller:0.14.18 secrets
  
127.0.0.1:44301/openstack/openstack-controller:0.14.18 secrets
 
127.0.0.1:44301/openstack/openstack-controller:0.14.18 health
  
127.0.0.1:44301/openstack/openstack-controller:0.14.18 health
Строка 125: Строка 128:
 
127.0.0.1:44301/openstack/openstack-controller:0.14.18 osdplsecret
  
127.0.0.1:44301/openstack/openstack-controller:0.14.18 osdplsecret
 
127.0.0.1:44301/openstack/openstack-controller:0.14.18 tf-secrets
  
127.0.0.1:44301/openstack/openstack-controller:0.14.18 tf-secrets
  +
End output for POD: openstack-controller-7896bc6d4d-l5mrz
   
−
openstack-controller-admission-5d78bf4d84-jmghf 127.0.0.1:44301/openstack/openstack-controller:0.14.18 openstack-operator
 +
Start output for POD: openstack-controller-admission-5d78bf4d84-jmghf
  +
Containes:
  +
127.0.0.1:44301/openstack/openstack-controller:0.14.18 openstack-operator
  +
End output for POD: openstack-controller-admission-5d78bf4d84-jmghf
   
−
openstack-controller-exporter-9f775b56f-gnntj 127.0.0.1:44301/openstack/openstack-controller:0.14.18 exporter
 +
Start output for POD: openstack-controller-exporter-9f775b56f-gnntj
  +
Containes:
  +
127.0.0.1:44301/openstack/openstack-controller:0.14.18 exporter
  +
End output for POD: openstack-controller-exporter-9f775b56f-gnntj
  +
</PRE>
  +
===Найти все поды по метке (<code>-l</code>) и вернуть список имеджей===
  +
<code>kubectl get pods --all-namespaces -o=jsonpath='{..image}' -l app=prometheus</code>
   
  +
===Число Pod на node===
  +
<PRE>
  +
kubectl get pod --all-namespaces -o json | jq '.items | group_by(.spec.nodeName) | map({"nodeName": .[0].spec.nodeName, "count": length}) | sort_by(.count)'
  +
</PRE>
  +
Пример результата
  +
<PRE>
  +
[
  +
...
  +
{
  +
"nodeName": "os-control-1",
  +
"count": 55
  +
},
  +
{
  +
"nodeName": "os-control-3",
  +
"count": 70
  +
}
  +
]
 
</PRE>
  
</PRE>
−
Get containers by Pod-Label (Return the container Image) kubectl get pods --all-namespaces -o=jsonpath="{..image}" -l app=prometheus
  
   
  +
===Список запущенных и не запущенных POD===
−
Pod count per node kubectl get po --all-namespaces -o json | jq '.items | group_by(.spec.nodeName) | map({"nodeName": .[0].spec.nodeName, "count": length}) | sort_by(.count)'
  
  +
<code>kubectl get pods --all-namespaces --field-selector=status.phase=Running</code>
  +
<PRE>
  +
NAMESPACE NAME READY STATUS RESTARTS AGE
  +
ceph-lcm-mirantis ceph-controller-567884bc4f-h6ppp 2/2 Running 0 4d1h
  +
ceph-lcm-mirantis ceph-disk-daemon-22cl2 1/1 Running 1 (4d17h ago) 26d
  +
ceph-lcm-mirantis ceph-disk-daemon-gcsnz 1/1 Running 1 (4d17h ago) 26d
  +
ceph-lcm-mirantis ceph-disk-daemon-wwvrk 1/1 Running 1 (4d17h ago) 26d
  +
ceph-lcm-mirantis ceph-maintenance-controller-76669864dc-njppm 1/1 Running 0 4d1h
  +
ceph-lcm-mirantis ceph-request-controller-7d7d9ffc77-jz6lq 3/3 Running 0 4d1h
  +
ceph-lcm-mirantis ceph-status-controller-556574b56b-bgh78 1/1 Running 0 4d1h
  +
kaas client-certificate-controller-59546996f-dqmhf 1/1 Running 0 4d1h
  +
...
  +
</PRE>
   
−
Get Running pods only kubectl get pods --all-namespaces --field-selector=status.phase=Running
 +
<code>kubectl get pods --all-namespaces --field-selector='status.phase!=Running</code>
  +
<PRE>
−  
  +
NAMESPACE NAME READY STATUS RESTARTS AGE
−
Get !Running pods kubectl get pods --all-namespaces --field-selector='status.phase!=Running
  
  +
openstack aodh-db-purge-28472041-2plwc 0/1 Completed 0 6h24m
−  
  +
openstack aodh-ks-user-pznnt 0/1 Completed 0 4d2h
−
Get all pods in All-namespaces, return mem and cpu (FOR EACH POD!)
  
  +
openstack barbican-db-purge-28471921-st487 0/1 Completed 0 8h
  +
openstack barbican-ks-user-k9c8s 0/1 Completed 0 4d2h
  +
openstack ceilometer-ks-user-kh7mh 0/1 Completed 0 4d2h
  +
...
  +
</PRE>
   
  +
===Get all pods in All-namespaces, return mem and cpu (FOR EACH POD!)===
  +
<PRE>
 
kubectl get po -n elasticsearch --selector=service=elasticsearch-data -o=jsonpath="{range .items[*]}{.metadata.namespace}:{.metadata.name}{'\n'}{range .spec.containers[*]} {.resources.requests.cpu}{'\t'}{end}{'\n'}{end}"
  
kubectl get po -n elasticsearch --selector=service=elasticsearch-data -o=jsonpath="{range .items[*]}{.metadata.namespace}:{.metadata.name}{'\n'}{range .spec.containers[*]} {.resources.requests.cpu}{'\t'}{end}{'\n'}{end}"
  +
</PRE>
−
Describe a POD and list its IP kubectl describe pod 'kube-dns-2948363707' -n kube-system | grep IP | sed -E 's/IP:[[:space:]]+//'
  
   
−
Get the external IPs of PODs kubectl get nodes --all-namespaces -o jsonpath='{.items[*].status.addresses[?(@.type=="ExternalIP")].address}'
 +
===Get the external IPs of PODs ===
  +
kubectl get nodes --all-namespaces -o jsonpath='{.items[*].status.addresses[?(@.type=="ExternalIP")].address}'
   
  +
===Show the images and their versions for each pod ===
−
Show the images and their versions for each pod kubectl get pods --all-namespaces -o jsonpath="{..image}" | tr -s '[[:space:]]' '\n' | sort | uniq -c kubectl get pods --all-namespaces -o jsonpath="{.items[*].spec.containers[*].image}" kubectl get pods -n default --field-selector=status.phase=Running --selector=app=my-service -o jsonpath="{..image}" | tr -s '[[:space:]]' '\n' | sort | uniq -c
  
  +
<PRE>
  +
kubectl get pods --all-namespaces -o jsonpath="{..image}" | tr -s '[[:space:]]' '\n' | sort | uniq -c kubectl get pods --all-namespaces -o jsonpath="{.items[*].spec.containers[*].image}" kubectl get pods -n default --field-selector=status.phase=Running --selector=app=my-service -o jsonpath="{..image}" | tr -s '[[:space:]]' '\n' | sort | uniq -c
  +
</PRE>
  +
===Delete terminated/evicted pods ===
  +
<PRE>
  +
kubectl get po --selector=app=devops-logstash-service --field-selector=status.phase=Failed --no-headers -n elasticsearch | awk '{ print $1 }' | xargs -I {} kubectl delete po -n elasticsearch {}
  +
</PRE>
   
  +
==Deployments==
−
Delete terminated/evicted pods kubectl get po --selector=app=devops-logstash-service --field-selector=status.phase=Failed --no-headers -n elasticsearch | awk '{ print $1 }' | xargs -I {} kubectl delete po -n elasticsearch {}
  
   
  +
Get deployments, output as JSON, format output with JQ
−
Deployments
  
  +
<PRE>
  +
kubectl get deployment --namespace=kube-system -o=json | jq '.items[].metadata.labels'
  +
</PRE>
  +
Get K8S Deployments matching the selector (look for key k8s-app with value heapster)
  +
<PRE>
  +
kubectl get deployment --all-namespaces --selector=k8s-app=heapster kubectl get svc --namespace=kube-system --selector=name=tiller
  +
</PRE>
   
  +
Scale a deployment
−
Get deployments, output as JSON, format output with JQ kubectl get deployment --namespace=kube-system -o=json | jq '.items[].metadata.labels'
  
  +
<PRE>
  +
kubectl scale -n kube-system --replicas=0 'deployments/kube-dns' kubectl scale -n kube-system --replicas=2 'deployments/kube-dns'
  +
</PRE>
  +
Export the YAML file of a deployment (for reuse, modification, etc,.)
  +
<PRE>
  +
kubectl get deployments ghost --export -n ghost -o yaml > ghost.yaml
  +
</PRE>
   
  +
==Replicasets==
−
Get K8S Deployments matching the selector (look for key k8s-app with value heapster) kubectl get deployment --all-namespaces --selector=k8s-app=heapster kubectl get svc --namespace=kube-system --selector=name=tiller
  
   
  +
===Get all replicasets sorted by the number of replicas===
−
Scale a deployment kubectl scale -n kube-system --replicas=0 'deployments/kube-dns' kubectl scale -n kube-system --replicas=2 'deployments/kube-dns'
  
  +
<PRE>
  +
kubectl get rs -n elasticsearch --sort-by='.spec.replicas'
  +
</PRE>
  +
===Get all replicasets where replicas==0 in a specific namespace===
  +
<PRE>
  +
kubectl get rs -n elasticsearch -o json | jq -r '.items[] |select(.status.replicas==0) | .metadata.name'
  +
</PRE>
  +
===Delete all replicasets that equal 0 in a specific namespace===
  +
<PRE>
  +
kubectl get rs -n elasticsearch -o json | jq -r '.items[] |select(.status.replicas==0) | .metadata.name' | xargs -I {} kubectl delete rs {}
  +
</PRE>
   
  +
==Services==
−
Export the YAML file of a deployment (for reuse, modification, etc,.) kubectl get deployments ghost --export -n ghost -o yaml > ghost.yaml
  
   
  +
===Get the self-link for a service using JSONPATH to format the output===
−
Replicasets
  
  +
<PRE>
  +
kubectl get svc heapster --namespace=kube-system -o=jsonpath={.metadata.selfLink}
  +
</PRE>
   
  +
===Get 'service' external LB address===
−
Get all replicasets sorted by the number of replicas kubectl get rs -n elasticsearch --sort-by='.spec.replicas'
  
  +
<PRE>
−  
−
Get all replicasets where replicas==0 in a specific namespace kubectl get rs -n elasticsearch -o json | jq -r '.items[] |select(.status.replicas==0) | .metadata.name'
 +
kubectl get svc --namespace=kube-system -o=json | jq -r '.items[] | .status.loadBalancer.ingress[0]'
  +
</PRE>
−  
  +
===List all services in cluster and their nodePorts===
−
Delete all replicasets that equal 0 in a specific namespace kubectl get rs -n elasticsearch -o json | jq -r '.items[] |select(.status.replicas==0) | .metadata.name' | xargs -I {} kubectl delete rs {}
  
  +
<PRE>
−  
  +
kubectl get --all-namespaces svc -o json | jq -r '.items[] | [.metadata.name,([.spec.ports[].nodePort | tostring ] | join("|"))] | @csv'
−
Services
  
  +
</PRE>
−  
−
Get the self-link for a service using JSONPATH to format the output kubectl get svc heapster --namespace=kube-system -o=jsonpath={.metadata.selfLink}
  
−  
−
Get 'service' external LB address kubectl get svc --namespace=kube-system -o=json | jq -r '.items[] | .status.loadBalancer.ingress[0]'
  
−  
−
List all services in cluster and their nodePorts: kubectl get --all-namespaces svc -o json | jq -r '.items[] | [.metadata.name,([.spec.ports[].nodePort | tostring ] | join("|"))] | @csv'
  
−  
−
Get Token for a service account & then get the secret
  
   
  +
===Get Token for a service account & then get the secret===
  +
<PRE>
 
DEFAULT_TOKEN=$(kubectl --namespace=kube-system get serviceaccount default -o jsonpath="{.secrets[0].name}")
  
DEFAULT_TOKEN=$(kubectl --namespace=kube-system get serviceaccount default -o jsonpath="{.secrets[0].name}")
  +
</PRE>
  +
<PRE>
 
TOKEN_VALUE=$(kubectl --namespace=kube-system get secret "$DEFAULT_TOKEN" -o go-template="{{.data.token}}" | base64 -d)
  
TOKEN_VALUE=$(kubectl --namespace=kube-system get secret "$DEFAULT_TOKEN" -o go-template="{{.data.token}}" | base64 -d)
  +
</PRE>
   
  +
==curl==
 
# Use it in a curl request to the api-server
  
# Use it in a curl request to the api-server
  +
<PRE>
−  
 
curl -k -H "Authorization: Bearer $TOKEN_VALUE" https://10.0.1.212:6443/version
  
curl -k -H "Authorization: Bearer $TOKEN_VALUE" https://10.0.1.212:6443/version
 
curl -k -H "Authorization: Bearer $TOKEN_VALUE" https://localhost:6443/version
  
curl -k -H "Authorization: Bearer $TOKEN_VALUE" https://localhost:6443/version
  +
</PRE>
−
Ingress
  
   
  +
==Ingress==
−
List all ingresses hosts and paths: kubectl get --all-namespaces ing -o='custom-columns=NAME:.metadata.name,HOSTS:.spec.rules[*].host,PATHS:.spec.rules[*].http.paths[*].path'
  
−
Persistent Volumes
  
   
  +
List all ingresses hosts and paths
−
List persistent volume claims in the elasticsearch namespace and for each print name & capacity kubectl get persistentvolumeclaims -n elasticsearch -o=jsonpath="{range .items[*]}{..name}{'\t'}{..capacity}{'\n'}{end}"
  
  +
<PRE>
  +
kubectl get --all-namespaces ing -o='custom-columns=NAME:.metadata.name,HOSTS:.spec.rules[*].host,PATHS:.spec.rules[*].http.paths[*].path'
  +
</PRE>
  +
  +
==Persistent Volumes==
  +
  +
List persistent volume claims in the elasticsearch namespace and for each print name & capacity
  +
<PRE>
  +
kubectl get persistentvolumeclaims -n elasticsearch -o=jsonpath="{range .items[*]}{..name}{'\t'}{..capacity}{'\n'}{end}"
  +
</PRE>

Текущая версия на 16:59, 19 февраля 2024

Шпаргалка по kubectl

много примеров тут: https://kubernetes.io/docs/reference/kubectl/quick-reference/

Частные случаи

--show-managed-fields - посмотреть какие были операции с объектом

иногда помогает найти когда были работы с объектом, например меняли ли его недавно (может не работать с CRD)
kubectl -n kaas get secret iam-api-secrets --show-managed-fields -o yaml

Вывод переносов строк с -o jsonpath= {"\n"}

kubectl get pods -A -o jsonpath='{range .items[*]}{.status.podIP}{"\n"}{end}'

Сортировка

kubectl get events --sort-by=.metadata.creationTimestamp

Поиск по полю

Скорее всего это все можно делать используя -o jsonpath но я встречал неоднократные упоминания о багах, по-тому использую jq

простое условие

Получить только имя репликасетов у которых число реплик = 1
kubectl get rs -o json | jq -r '.items[] |select(.status.replicas==1) | .metadata.name'

Поиск специфичного объекта

Получить секрет(ы) у которых в аннотации в поле .metadata.annotations."lcm.mirantis.com/cluster-secret-type" значение "kubeconfig"
(Пример специфичен для окружений определенного вида, но работает для любых объектов) kubectl get secret -A -o json | jq ' .items[] | select(.metadata.annotations."lcm.mirantis.com/cluster-secret-type"=="kubeconfig") | .metadata.name'

Как писать такие запросы:

  • Получить список, тут видно что все объекты помешены в List .items[]
{
  "apiVersion": "v1",
  "items": [
    {
      "apiVersion": "v1",
      "kind": "Secret",
      "metadata": {
        "annotations": {
          "kubernetes.io/service-account.name": "default",
          "kubernetes.io/service-account.uid": "1ccf69fb-2f37-4684-bb16-19a8f20fd2fe"
        },
  • Конструкция jq '.items[] | <команда>' проверяет каждый элемент списка, возвращая только соответствующие условию (в примере это условие select(.metadata.annotations."lcm.mirantis.com/cluster-secret-type"=="kubeconfig"))
  • .metadata.name - вывести имя (но тут можно и содержимое или другое поле по желанию)


Nodes

Метки

kubectl get nodes --show-labels

Top usage

kubectl top nodes

Фильтрация вывода - все кроме

kubectl get node --selector='!node-role.kubernetes.io/master'

Без хедеров

kubectl get node --no-headers -o custom-columns=NAME:.metadata.name

Сортировка

kubectl get node --sort-by .metadata.creationTimestamp

Получение определенных полей с переименованием по аналогии с SELECT AS ..

kubectl get no -o json | jq -r '[.items[] | {name:.metadata.name, id:.spec.externalID, unschedulable:.spec.unschedulable}]' 

[
...
  {
    "name": "storage-3",
    "id": null,
    "unschedulable": null
  }
]

Простой просмотр утилизации

Это конечно хорошо бы улучшить kubectl describe node | grep -A5 "Allocated" 

Allocated resources:
  (Total limits may be over 100 percent, i.e., overcommitted.)
  Resource           Requests     Limits
  --------           --------     ------
  cpu                1020m (8%)   4100m (34%)
  memory             1178Mi (0%)  3444Mi (2%)


Pods

Список контейнеров в поде

Для каждого контейнера pod-name, container-image container-name
Довольно коряво, конечно, но можно использовать как основу для своих скриптов 


kubectl \
    get pods \
    -n osh-system \
    --selector=app.kubernetes.io/instance=openstack-operator \
    -o=jsonpath='{range .items[*]}{"\nStart output for POD: "}{.metadata.name}{"\nContaines:\n"}{range .spec.containers[*]}{.image}{"\t"}{.name}{"\n"}{end}{"End output for POD: "}{.metadata.name}{"\n"}{end}'

Пример вывода: 

Start output for POD: openstack-controller-7896bc6d4d-l5mrz
Containes:
127.0.0.1:44301/openstack/openstack-controller:0.14.18	osdpl
127.0.0.1:44301/openstack/openstack-controller:0.14.18	secrets
127.0.0.1:44301/openstack/openstack-controller:0.14.18	health
127.0.0.1:44301/openstack/openstack-controller:0.14.18	node
127.0.0.1:44301/openstack/openstack-controller:0.14.18	nodemaintenancerequest
127.0.0.1:44301/openstack/openstack-controller:0.14.18	ceph-secrets
127.0.0.1:44301/openstack/openstack-controller:0.14.18	osdplstatus
127.0.0.1:44301/openstack/openstack-controller:0.14.18	osdplsecret
127.0.0.1:44301/openstack/openstack-controller:0.14.18	tf-secrets
End output for POD: openstack-controller-7896bc6d4d-l5mrz

Start output for POD: openstack-controller-admission-5d78bf4d84-jmghf
Containes:
127.0.0.1:44301/openstack/openstack-controller:0.14.18	openstack-operator
End output for POD: openstack-controller-admission-5d78bf4d84-jmghf

Start output for POD: openstack-controller-exporter-9f775b56f-gnntj
Containes:
127.0.0.1:44301/openstack/openstack-controller:0.14.18	exporter
End output for POD: openstack-controller-exporter-9f775b56f-gnntj

Найти все поды по метке (-l) и вернуть список имеджей

kubectl get pods --all-namespaces -o=jsonpath='{..image}' -l app=prometheus

Число Pod на node

kubectl get pod --all-namespaces -o json | jq '.items | group_by(.spec.nodeName) | map({"nodeName": .[0].spec.nodeName, "count": length}) | sort_by(.count)'

Пример результата 

[
  ...
  {
    "nodeName": "os-control-1",
    "count": 55
  },
  {
    "nodeName": "os-control-3",
    "count": 70
  }
]

Список запущенных и не запущенных POD

kubectl get pods --all-namespaces --field-selector=status.phase=Running 

NAMESPACE                NAME                                                   READY   STATUS    RESTARTS         AGE
ceph-lcm-mirantis        ceph-controller-567884bc4f-h6ppp                       2/2     Running   0                4d1h
ceph-lcm-mirantis        ceph-disk-daemon-22cl2                                 1/1     Running   1 (4d17h ago)    26d
ceph-lcm-mirantis        ceph-disk-daemon-gcsnz                                 1/1     Running   1 (4d17h ago)    26d
ceph-lcm-mirantis        ceph-disk-daemon-wwvrk                                 1/1     Running   1 (4d17h ago)    26d
ceph-lcm-mirantis        ceph-maintenance-controller-76669864dc-njppm           1/1     Running   0                4d1h
ceph-lcm-mirantis        ceph-request-controller-7d7d9ffc77-jz6lq               3/3     Running   0                4d1h
ceph-lcm-mirantis        ceph-status-controller-556574b56b-bgh78                1/1     Running   0                4d1h
kaas                     client-certificate-controller-59546996f-dqmhf          1/1     Running   0                4d1h
...

kubectl get pods --all-namespaces --field-selector='status.phase!=Running 

NAMESPACE    NAME                                       READY   STATUS      RESTARTS   AGE
openstack    aodh-db-purge-28472041-2plwc               0/1     Completed   0          6h24m
openstack    aodh-ks-user-pznnt                         0/1     Completed   0          4d2h
openstack    barbican-db-purge-28471921-st487           0/1     Completed   0          8h
openstack    barbican-ks-user-k9c8s                     0/1     Completed   0          4d2h
openstack    ceilometer-ks-user-kh7mh                   0/1     Completed   0          4d2h
...

Get all pods in All-namespaces, return mem and cpu (FOR EACH POD!)

kubectl get po -n elasticsearch --selector=service=elasticsearch-data -o=jsonpath="{range .items[*]}{.metadata.namespace}:{.metadata.name}{'\n'}{range .spec.containers[*]} {.resources.requests.cpu}{'\t'}{end}{'\n'}{end}"

Get the external IPs of PODs

kubectl get nodes --all-namespaces -o jsonpath='{.items[*].status.addresses[?(@.type=="ExternalIP")].address}'

Show the images and their versions for each pod

kubectl get pods --all-namespaces -o jsonpath="{..image}" | tr -s '[[:space:]]' '\n' | sort | uniq -c kubectl get pods --all-namespaces -o jsonpath="{.items[*].spec.containers[*].image}" kubectl get pods -n default --field-selector=status.phase=Running --selector=app=my-service -o jsonpath="{..image}" | tr -s '[[:space:]]' '\n' | sort | uniq -c

Delete terminated/evicted pods

kubectl get po --selector=app=devops-logstash-service --field-selector=status.phase=Failed --no-headers -n elasticsearch | awk '{ print $1 }' | xargs -I {} kubectl delete po -n elasticsearch {}

Deployments

Get deployments, output as JSON, format output with JQ 

kubectl get deployment --namespace=kube-system -o=json | jq '.items[].metadata.labels'

Get K8S Deployments matching the selector (look for key k8s-app with value heapster) 

kubectl get deployment --all-namespaces --selector=k8s-app=heapster kubectl get svc --namespace=kube-system --selector=name=tiller

Scale a deployment 

kubectl scale -n kube-system --replicas=0 'deployments/kube-dns' kubectl scale -n kube-system --replicas=2 'deployments/kube-dns'

Export the YAML file of a deployment (for reuse, modification, etc,.) 

kubectl get deployments ghost --export -n ghost -o yaml > ghost.yaml

Replicasets

Get all replicasets sorted by the number of replicas

kubectl get rs -n elasticsearch --sort-by='.spec.replicas'

Get all replicasets where replicas==0 in a specific namespace

kubectl get rs -n elasticsearch -o json | jq -r '.items[] |select(.status.replicas==0) | .metadata.name'

Delete all replicasets that equal 0 in a specific namespace

kubectl get rs -n elasticsearch -o json | jq -r '.items[] |select(.status.replicas==0) | .metadata.name' | xargs -I {} kubectl delete rs {}

Services

Get the self-link for a service using JSONPATH to format the output

kubectl get svc heapster --namespace=kube-system -o=jsonpath={.metadata.selfLink}

Get 'service' external LB address

kubectl get svc --namespace=kube-system -o=json | jq -r '.items[] | .status.loadBalancer.ingress[0]'

List all services in cluster and their nodePorts

kubectl get --all-namespaces svc -o json | jq -r '.items[] | [.metadata.name,([.spec.ports[].nodePort | tostring ] | join("|"))] | @csv'

Get Token for a service account & then get the secret

DEFAULT_TOKEN=$(kubectl --namespace=kube-system get serviceaccount default -o jsonpath="{.secrets[0].name}")

TOKEN_VALUE=$(kubectl --namespace=kube-system get secret "$DEFAULT_TOKEN" -o go-template="{{.data.token}}" | base64 -d)

curl

  1. Use it in a curl request to the api-server
curl -k -H "Authorization: Bearer $TOKEN_VALUE" https://10.0.1.212:6443/version
curl -k -H "Authorization: Bearer $TOKEN_VALUE" https://localhost:6443/version

Ingress

List all ingresses hosts and paths 

kubectl get --all-namespaces ing -o='custom-columns=NAME:.metadata.name,HOSTS:.spec.rules[*].host,PATHS:.spec.rules[*].http.paths[*].path'

Persistent Volumes

List persistent volume claims in the elasticsearch namespace and for each print name & capacity 

kubectl get persistentvolumeclaims -n elasticsearch -o=jsonpath="{range .items[*]}{..name}{'\t'}{..capacity}{'\n'}{end}"
Источник — https://noname.com.ua/mediawiki/index.php?title=Kubectl&oldid=12696