资源池升级
文档版本
| 序号 | 内容 | 更新人 | 更新日期 | 版本 |
|---|---|---|---|---|
| 1 | 文档初始化 | 许向 | 2018-5-13 | 0.1 |
| 2 | 文档补全 | 许向 | 2019-2-11 | 0.2 |
资源池扩容1
在某些情况下资源节点的配置并不能满足实际应用需求，比如某个应用需要2颗CPU，但是物理节点只有1颗CPU。GCP的扩容只能在同级别的主机资源上，也就是说1Core的机器再扩容也是1Core * n，这个时候只能通过升级底层资源池的配置来达到扩容的需求
显示当前pool
$ gcloud container node-pools list --cluster cluster-1 --zone asia-northeast1-a NAME MACHINE_TYPE DISK_SIZE_GB NODE_VERSION default-pool n1-standard-1 100 1.9.3-gke.0创建一个大号的pool
$ gcloud container node-pools create n1-standard-2 --cluster cluster-1 --zone asia-northeast1-a --machine-type=n1-standard-2 --num-nodes=2 Creating node pool n1-standard-2...done. Created [https://container.googleapis.com/v1/projects/demo-1/zones/asia-northeast1-a/clusters/cluster-1/nodePools/n1-standard-2]. NAME MACHINE_TYPE DISK_SIZE_GB NODE_VERSION n1-standard-2 n1-standard-2 100 1.9.3-gke.0检查
$ gcloud container node-pools list --cluster cluster-1 --zone asia-northeast1-a NAME MACHINE_TYPE DISK_SIZE_GB NODE_VERSION default-pool n1-standard-1 100 1.9.3-gke.0 n1-standard-2 n1-standard-2 100 1.9.3-gke.0$ kubectl get nodes NAME STATUS ROLES AGE VERSION gke-cluster-1-default-pool-ddf96b65-3f5f Ready <none> 12d v1.9.3-gke.0 gke-cluster-1-default-pool-ddf96b65-drpv Ready <none> 10d v1.9.3-gke.0 gke-cluster-1-default-pool-ddf96b65-ggg3 Ready <none> 12d v1.9.3-gke.0 gke-cluster-1-n1-standard-2-92d50459-tbv0 Ready <none> 6m v1.9.3-gke.0 gke-cluster-1-n1-standard-2-92d50459-tlkz Ready <none> 6m v1.9.3-gke.0$ kubectl get pods -o=wide NAME READY STATUS RESTARTS AGE IP NODE cluster-devel-service-os-1-675c6d95d7-nqtnt 1/1 Running 0 22m 10.32.4.89 gke-cluster-1-default-pool-ddf96b65-drpv cluster-prod-app-api-1-c7d7d9c79-9b8v6 2/2 Running 0 22m 10.32.2.44 gke-cluster-1-default-pool-ddf96b65-3f5f cluster-prod-app-chat-1-7dcdd4575c-8gs9q 1/1 Running 0 5d 10.32.0.60 gke-cluster-1-default-pool-ddf96b65-ggg3 cluster-prod-app-daemons-1-566c6f5f58-rhrrc 2/2 Running 0 1h 10.32.4.87 gke-cluster-1-default-pool-ddf96b65-drpv cluster-prod-app-static-1-667bfb756f-kns9x 1/1 Running 2 1d 10.32.4.68 gke-cluster-1-default-pool-ddf96b65-drpv cluster-prod-service-geth-1-5b54bcf9bc-rmsr9 1/1 Running 0 11m 10.32.0.73 gke-cluster-1-default-pool-ddf96b65-ggg3 cluster-prod-service-lb-1-9f68489dc-j5h2c 1/1 Running 0 2d 10.32.0.65 gke-cluster-1-default-pool-ddf96b65-ggg3 cluster-prod-service-redis-1-66c9cdb5fd-jlctd 1/1 Running 0 11d 10.32.0.33 gke-cluster-1-default-pool-ddf96b65-ggg3 cluster-prod-service-rmq-1-6b8848dccb-xn64n 1/1 Running 0 5d 10.32.0.61 gke-cluster-1-default-pool-ddf96b65-ggg3迁移工作负载
此步骤又分成2个小步骤
- Cordon the existing node pool
Drain the existing node pool
Cordon
首先检查当前default-pool
$ kubectl get nodes -l cloud.google.com/gke-nodepool=default-pool NAME STATUS ROLES AGE VERSION gke-cluster-1-default-pool-ddf96b65-3f5f Ready <none> 12d v1.9.3-gke.0 gke-cluster-1-default-pool-ddf96b65-drpv Ready <none> 10d v1.9.3-gke.0 gke-cluster-1-default-pool-ddf96b65-ggg3 Ready <none> 12d v1.9.3-gke.0cordon
$ for node in $(kubectl get nodes -l cloud.google.com/gke-nodepool=default-pool -o=name); do kubectl cordon "$node"; done node "gke-cluster-1-default-pool-ddf96b65-3f5f" cordoned node "gke-cluster-1-default-pool-ddf96b65-drpv" cordoned node "gke-cluster-1-default-pool-ddf96b65-ggg3" cordoned查看
$ kubectl get nodes NAME STATUS ROLES AGE VERSION gke-cluster-1-default-pool-ddf96b65-3f5f Ready,SchedulingDisabled <none> 13d v1.9.3-gke.0 gke-cluster-1-default-pool-ddf96b65-drpv Ready,SchedulingDisabled <none> 10d v1.9.3-gke.0 gke-cluster-1-default-pool-ddf96b65-ggg3 Ready,SchedulingDisabled <none> 13d v1.9.3-gke.0 gke-cluster-1-n1-standard-2-92d50459-tbv0 Ready <none> 55m v1.9.3-gke.0 gke-cluster-1-n1-standard-2-92d50459-tlkz Ready <none> 55m v1.9.3-gke.0Drain
$ for node in $(kubectl get nodes -l cloud.google.com/gke-nodepool=default-pool -o=name); do kubectl drain --force --ignore-daemonsets --delete-local-data --grace-period=10 "$node"; done WARNING: Deleting pods not managed by ReplicationController, ReplicaSet, Job, DaemonSet or StatefulSet: kube-proxy-gke-cluster-1-default-pool-ddf96b65-3f5f; Deleting pods with local storage: cluster-prod-app-api-1-c7d7d9c79-9b8v6; Ignoring DaemonSet-managed pods: fluentd-gcp-v2.0.10-vkzvd pod "kube-dns-autoscaler-69c5cbdcdd-tmzq8" evicted pod "kube-dns-5c5884448b-lpkrf" evicted pod "kube-dns-5c5884448b-f6gsz" evicted pod "event-exporter-v0.1.7-7c4f8bb746-47rvr" evicted pod "cluster-prod-app-api-1-c7d7d9c79-9b8v6" evicted node "gke-cluster-1-default-pool-ddf96b65-3f5f" drained node "gke-cluster-1-default-pool-ddf96b65-drpv" already cordoned WARNING: Ignoring DaemonSet-managed pods: fluentd-gcp-v2.0.10-vc862; Deleting pods not managed by ReplicationController, ReplicaSet, Job, DaemonSet or StatefulSet: kube-proxy-gke-cluster-1-default-pool-ddf96b65-drpv; Deleting pods with local storage: cluster-prod-app-daemons-1-566c6f5f58-rhrrc pod "cluster-prod-service-geth-1-5b54bcf9bc-zqs5p" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-4rcq6" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-zhjz4" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-tj4wc" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-f2rcw" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-xjkcm" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-rl4w9" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-vv56l" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-xvzrz" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-8949p" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-4pvk2" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-5rq27" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-bdtfd" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-v5gwq" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-jlvcg" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-rvqm6" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-g9zvr" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-mt7b7" evicted pod "cluster-devel-service-os-1-675c6d95d7-nqtnt" evicted pod "heapster-v1.5.0-6d8547cdc9-5bn5w" evicted pod "cluster-prod-app-static-1-667bfb756f-kns9x" evicted pod "cluster-prod-app-daemons-1-566c6f5f58-rhrrc" evicted node "gke-cluster-1-default-pool-ddf96b65-drpv" drained node "gke-cluster-1-default-pool-ddf96b65-ggg3" already cordoned WARNING: Ignoring DaemonSet-managed pods: fluentd-gcp-v2.0.10-mw8n5; Deleting pods not managed by ReplicationController, ReplicaSet, Job, DaemonSet or StatefulSet: kube-proxy-gke-cluster-1-default-pool-ddf96b65-ggg3; Deleting pods with local storage: kubernetes-dashboard-775fc9968-9ntzq pod "cluster-prod-service-geth-1-5b54bcf9bc-rmsr9" evicted pod "cluster-prod-service-geth-1-5b54bcf9bc-wj25v" evicted pod "l7-default-backend-57856c5f55-46psz" evicted pod "cluster-prod-service-redis-1-66c9cdb5fd-jlctd" evicted pod "cluster-prod-app-chat-1-7dcdd4575c-8gs9q" evicted pod "kubernetes-dashboard-775fc9968-9ntzq" evicted pod "metrics-server-v0.2.1-7f8dd98c8f-vvcm8" evicted pod "cluster-prod-service-lb-1-9f68489dc-j5h2c" evicted pod "cluster-prod-service-rmq-1-6b8848dccb-xn64n" evicted node "gke-cluster-1-default-pool-ddf96b65-ggg3" drained查看
$ kubectl get pods -o wide NAME READY STATUS RESTARTS AGE IP NODE cluster-devel-service-os-1-675c6d95d7-4b6bc 1/1 Running 0 3m 10.32.10.7 gke-cluster-1-n1-standard-2-92d50459-tlkz cluster-prod-app-api-1-c7d7d9c79-bzgkx 2/2 Running 0 4m 10.32.9.4 gke-cluster-1-n1-standard-2-92d50459-tbv0 cluster-prod-app-chat-1-7dcdd4575c-zpqdl 1/1 Running 0 2m 10.32.10.10 gke-cluster-1-n1-standard-2-92d50459-tlkz cluster-prod-app-daemons-1-566c6f5f58-hhgxb 2/2 Running 0 3m 10.32.10.8 gke-cluster-1-n1-standard-2-92d50459-tlkz cluster-prod-app-static-1-667bfb756f-zjf68 1/1 Running 0 3m 10.32.10.6 gke-cluster-1-n1-standard-2-92d50459-tlkz cluster-prod-service-geth-1-5b54bcf9bc-xdd5p 1/1 Running 0 48m 10.32.10.3 gke-cluster-1-n1-standard-2-92d50459-tlkz cluster-prod-service-lb-1-9f68489dc-n8fsd 1/1 Running 0 2m 10.32.10.9 gke-cluster-1-n1-standard-2-92d50459-tlkz cluster-prod-service-redis-1-66c9cdb5fd-k54zb 1/1 Running 0 2m 10.32.10.11 gke-cluster-1-n1-standard-2-92d50459-tlkz cluster-prod-service-rmq-1-6b8848dccb-kbzf8 1/1 Running 0 2m 10.32.9.8 gke-cluster-1-n1-standard-2-92d50459-tbv0
移除老pool
$ gcloud container node-pools delete default-pool --cluster cluster-1 --zone asia-northeast1-a The following node pool will be deleted. [default-pool] in cluster [cluster-1] in [asia-northeast1-a] Do you want to continue (Y/n)? y Deleting node pool default-pool...done. Deleted [https://container.googleapis.com/v1/projects/demo-1/zones/asia-northeast1-a/clusters/cluster-1/nodePools/default-pool].查看
$ gcloud container node-pools list --cluster cluster-1 --zone asia-northeast1-a NAME MACHINE_TYPE DISK_SIZE_GB NODE_VERSION n1-standard-2 n1-standard-2 100 1.9.3-gke.0
参考文档
1. migrating-node-pool ↩