Kubectl Foreach 在多个集群中执行 Kubectl 命令

    作者:匿名更新于: 2022-12-03 20:41:42

      今天偶然间发现了一个 kubectl 插件 kubectl foreach​ ,可以在多个集群(contexts​)上执行 kubectl 命令。比如 kubectl foreach cluster-1 cluster-2 -- get po -n kube-system 。

      上周在写 K8s 多集群的流量调度 的 demo 部分时需要不停地在多个集群中安装组件、部署应用,或者执行各种命令。当时是通过 Linux shell 脚本并通过工具 kubectx 进行集群的切换,像这样:

      或者这样:

      操作繁琐,很是痛苦。

      今天偶然间发现了一个 kubectl 插件 kubectl foreach​ ,可以在多个集群(contexts​)上执行 kubectl 命令。比如 kubectl foreach cluster-1 cluster-2 -- get po -n kube-system 。

      插件安装和使用很简单,通过 krew 进行安装:

      复制

      1.  kubectl krew install foreach

      使用也很简单:

      复制

      1.  kubectl foreach -h

      2.  Usage:

      3.  kubectl foreach [OPTIONS] [PATTERN]... -- [KUBECTL_ARGS...]

          4.

      5.  Patterns can be used to match context names from kubeconfig:

      6.  (empty): matches all contexts

      7.  NAME: matches context with exact name

      8.  /PATTERN/: matches context with regular expression

      9.  ^NAME: remove context with exact name from the matched results

      10.  ^/PATTERN/: remove contexts matching the regular expression from the results

          11.

      12.  Options:

      13.  -c=NUM Limit parallel executions (default: 0, unlimited)

      14.  -I=VAL Replace VAL occurring in KUBECTL_ARGS with context name

      15.  -q Disable and accept confirmation prompts ($KUBECTL_FOREACH_DISABLE_PROMPTS)

      16.  -h/--help Print help

          17.

      18.  Examples:

      19.  # get nodes on contexts named a b c

      20.  kubectl foreach a b c -- get nodes

          21.

      22.  # get nodes on all contexts named c0..9 except c1 (note the escaping)

      23.  kubectl foreach '/^c[0-9]/' ^c1 -- get nodes

          24.

      25.  # get nodes on all contexts that has "prod" but not "foo"

      26.  kubectl foreach /prod/ ^/foo/ -- get nodes

          27.

      28.  # use 'kubectl tail' plugin to follow logs of pods in contexts named *test*

      29.  kubectl foreach -I _ /test/ -- tail --cnotallow=_ -l app=foo

      接下来测试下,使用 k3d 创建 3 个集群 (k3d 貌似不支持同时创建多个集群,还是需要 for 脚本来操作):

      复制

      1.  for CLUSTER_NAME in cluster-1 cluster-2 cluster-3

      2.  do

      3.  k3d cluster create ${CLUSTER_NAME} \\

      4.  --image docker.io/rancher/k3s:v1.23.8-k3s2 \\

      5.  --servers-memory 4g \\

      6.  --k3s-arg "--disable=traefik@server:0" \\

      7.  --no-lb \\

      8.  --timeout 120s \\

      9.  --wait

      10.  done

      集群安装完成:

      复制

      1.  k3d cluster list

      2.  NAME SERVERS AGENTS LOADBALANCER

      3.  cluster-1 1/1 0/0 false

      4.  cluster-2 1/1 0/0 false

      5.  cluster-3 1/1 0/0 false

      注意,k3d 安装的集群的 context 都带有前缀 k3d-​ ,在使用 kubectl foreach 的时候要注意:

      复制

      1.  kubectx

      2.  k3d-cluster-1

      3.  k3d-cluster-2

      4.  k3d-cluster-3

      比如查看各个集群中的 kube-sysmte 下的 pod:

      复制

      1.  kubectl foreach -q k3d-cluster-1 k3d-cluster-2 k3d-cluster-3 -- get po -n kube-system

      或者试试创建 deployment,这次我们不列出完整的 context name,而是使用正则 /cluster/:

      复制

      1.  kubectl foreach -q /cluster/ -- create deploy pipy --image flomesh/pipy -n default

      2.  Will run command in context(s):

      3.  - k3d-cluster-1

      4.  - k3d-cluster-2

      5.  - k3d-cluster-3

      6.  k3d-cluster-1 | deployment.apps/pipy created

      7.  k3d-cluster-3 | deployment.apps/pipy created

      8.  k3d-cluster-2 | deployment.apps/pipy created

      然后查看下 pod:

      复制

      1.  kubectl foreach -q /cluster/ -- get pod -n default

      2.  Will run command in context(s):

      3.  - k3d-cluster-1

      4.  - k3d-cluster-2

      5.  - k3d-cluster-3

      6.  k3d-cluster-1 | NAME READY STATUS RESTARTS AGE

      7.  k3d-cluster-1 | pipy-df659b55f-bnr27 1/1 Running 0 25s

      8.  k3d-cluster-3 | NAME READY STATUS RESTARTS AGE

      9.  k3d-cluster-3 | pipy-df659b55f-p9j49 1/1 Running 0 25s

      10.  k3d-cluster-2 | NAME READY STATUS RESTARTS AGE

      11.  k3d-cluster-2 | pipy-df659b55f-9bjgf 1/1 Running 0 25s

      查看日志:

      复制

      1.  kubectl foreach -q /cluster/ -- logs -l app=pipy -n default --tail 3

      2.  Will run command in context(s):

      3.  - k3d-cluster-1

      4.  - k3d-cluster-2

      5.  - k3d-cluster-3

      6.  k3d-cluster-2 | 2022-11-30 10:40:56.520 [INF] [listener] Listening on TCP port 8080 at 0.0.0.0

      7.  k3d-cluster-2 | 2022-11-30 10:40:56.520 [INF] [listener] Listening on TCP port 8081 at 0.0.0.0

      8.  k3d-cluster-2 | 2022-11-30 10:40:56.520 [INF] [listener] Listening on TCP port 8082 at 0.0.0.0

      9.  k3d-cluster-1 | 2022-11-30 10:40:56.551 [INF] [listener] Listening on TCP port 8080 at 0.0.0.0

      10.  k3d-cluster-1 | 2022-11-30 10:40:56.551 [INF] [listener] Listening on TCP port 8081 at 0.0.0.0

      11.  k3d-cluster-1 | 2022-11-30 10:40:56.551 [INF] [listener] Listening on TCP port 8082 at 0.0.0.0

      12.  k3d-cluster-3 | 2022-11-30 10:40:55.813 [INF] [listener] Listening on TCP port 8080 at 0.0.0.0

      13.  k3d-cluster-3 | 2022-11-30 10:40:55.813 [INF] [listener] Listening on TCP port 8081 at 0.0.0.0

      14.  k3d-cluster-3 | 2022-11-30 10:40:55.813 [INF] [listener] Listening on TCP port 8082 at 0.0.0.0

      注意,多集群的操作要谨慎,尤其是使用正则来匹配 context name;还有 ​-q 参数会跳过要操作的集群提醒,直接执行命令。​

      来源: 云原生指北

        >>>>>>点击进入计算专题

云计算 更多推荐

课课家教育

未登录

1