Nesta página, fornecemos instruções sobre como atualizar dinamicamente os rótulos e taints de nó de um pool de nós em um cluster padrão do Google Kubernetes Engine (GKE).
Os taints e rótulos de nós estão nos metadados de objeto dos nós do Kubernetes. Os rótulos são usados para programar pods em nós específicos, em que os taints podem ser usados para afastar os pods deles.
Com a API Kubernetes Engine, é possível aplicar atualizações nos rótulos de nós e taints de nó de um pool de nós atual do GKE sem precisar recriar o pool de nós ou interromper a execução de cargas de trabalho. A configuração do pool de nós atualizada é preservada no GKE. Assim, os upgrades futuros do pool de nós e as novas provisões de nós no pool usarão a nova configuração.
Limitações
Há algumas limitações no uso da API Kubernetes Engine para atualizar dinamicamente as configurações do pool de nós:
A versão do pool de nós precisa ser 1.19.7-gke.1500 ou posterior.
A versão do plano de controle do cluster precisa ser 1.23.4-gke.300 ou posterior para aplicar atualizações aos rótulos ou taints de nó para pools de nós atuais com o escalonador automático de clusters ativado.
Para clusters em versões anteriores, use a seguinte solução alternativa:
Desativar o escalonamento automático
no pool de nós e atualizar os rótulos e/ou taints do nó. Depois que as
atualizações forem aplicadas, reative o escalonamento automático.
Como atualizar rótulos de nós
Para atualizar os rótulos de um pool de nós, use o seguinte comando:
[[["Fácil de entender","easyToUnderstand","thumb-up"],["Meu problema foi resolvido","solvedMyProblem","thumb-up"],["Outro","otherUp","thumb-up"]],[["Difícil de entender","hardToUnderstand","thumb-down"],["Informações incorretas ou exemplo de código","incorrectInformationOrSampleCode","thumb-down"],["Não contém as informações/amostras de que eu preciso","missingTheInformationSamplesINeed","thumb-down"],["Problema na tradução","translationIssue","thumb-down"],["Outro","otherDown","thumb-down"]],["Última atualização 2025-04-03 UTC."],[],[],null,["# Update node labels and taints for existing node pools\n\n[Standard](/kubernetes-engine/docs/concepts/choose-cluster-mode)\n\n*** ** * ** ***\n\nThis page provides instructions on how to dynamically update the\nnode labels and node taints of an existing node pool in a\nGoogle Kubernetes Engine (GKE) Standard cluster.\n\nYou can also dynamically apply and update [network tags](/vpc/docs/add-remove-network-tags)\nin GKE Autopilot or Standard clusters. For\ninstructions, refer to [Use network tags to apply firewall rules to nodes](/kubernetes-engine/docs/how-to/autopilot-network-tags).\n\nOverview\n--------\n\nNode [taints](https://kubernetes.io/docs/concepts/scheduling-eviction/taint-and-toleration/)\nand [labels](https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/)\nare in the object metadata of Kubernetes nodes. Labels are used to schedule Pods\non particular nodes, where taints can be used to steer Pods away from them.\n\nUsing the GKE API, you can apply updates on the node\nlabels, and node taints of an existing GKE node pool without\nneeding to recreate the node pool or disrupt running workloads. The updated node\npool configuration is preserved in GKE, so that future node pool\nupgrades and new node provisions in the node pool will use the new configuration.\n| **Warning:** Dynamic configuration changes are limited to these fields: resource manager tags, network tags, node labels and node taints. Any other field changes in the [UpdateNodePool\n| API](/kubernetes-engine/docs/reference/rest/v1/projects.locations.clusters.nodePools/update) will not occur dynamically, and will result in node recreation.\n\nLimitations\n-----------\n\nThere are some limitations for using the GKE API to dynamically\nupdate node pool configurations:\n\n- The version for the node pool must be 1.19.7-gke.1500 or later.\n- The version for the cluster's control plane must be 1.23.4-gke.300 or later to apply updates to node labels or node taints for existing node pools with [cluster autoscaler enabled](/kubernetes-engine/docs/how-to/cluster-autoscaler). For clusters on earlier versions, you can use the following workaround: [Disable autoscaling](/kubernetes-engine/docs/how-to/cluster-autoscaler#disable_autoscaling) on the node pool, and then update the node labels and/or taints. After the updates have been applied, [re-enable autoscaling](/kubernetes-engine/docs/how-to/cluster-autoscaler#enabling_autoscaling_for_an_existing_node_pool).\n\nUpdating node labels\n--------------------\n\nTo update node labels for a node pool, use the following command: \n\n gcloud container node-pools update \u003cvar translate=\"no\"\u003eNODEPOOL_NAME\u003c/var\u003e \\\n --node-labels=\u003cvar translate=\"no\"\u003e[\u003c/var\u003eNODE_LABEL,...] \\\n --cluster=\u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e\n --location=\u003cvar translate=\"no\"\u003eCONTROL_PLANE_LOCATION\u003c/var\u003e\n [GCLOUD_WIDE_FLAG ...]\n\n| **Warning:** This update overwrites any previous user-specified values.\n\nReplace the following:\n\n- \u003cvar translate=\"no\"\u003eNODEPOOL_NAME\u003c/var\u003e: the name of the node pool to update.\n- \u003cvar translate=\"no\"\u003e[NODE_LABEL,...]\u003c/var\u003e: the desired new node labels (for example, `label1=value1,label2=value2`).\n- \u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e: the name of the cluster.\n- \u003cvar translate=\"no\"\u003eCONTROL_PLANE_LOCATION\u003c/var\u003e: the Compute Engine [location](/compute/docs/regions-zones#available) of the control plane of your cluster. Provide a region for regional clusters, or a zone for zonal clusters.\n\nUpdating node taints\n--------------------\n\nTo update node taints for a node pool, use the following command: \n\n gcloud container node-pools update \u003cvar translate=\"no\"\u003eNODEPOOL_NAME\u003c/var\u003e \\\n --node-taints=\u003cvar translate=\"no\"\u003e[\u003c/var\u003eNODE_TAINT,...] \\\n --cluster=\u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e\n --location=\u003cvar translate=\"no\"\u003eCONTROL_PLANE_LOCATION\u003c/var\u003e\n [GCLOUD_WIDE_FLAG ...]\n\n| **Warning:** This update overwrites any previous user-specified values.\n\nReplace the following:\n\n- \u003cvar translate=\"no\"\u003eNODEPOOL_NAME\u003c/var\u003e: the name of the node pool to update.\n- \u003cvar translate=\"no\"\u003e[NODE_TAINT,...]\u003c/var\u003e: the desired new node taints (for example, `key1=val1:NoSchedule,key2=val2:PreferNoSchedule`).\n- \u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e: the name of the cluster.\n- \u003cvar translate=\"no\"\u003eCONTROL_PLANE_LOCATION\u003c/var\u003e: the Compute Engine [location](/compute/docs/regions-zones#available) of the control plane of your cluster. Provide a region for regional clusters, or a zone for zonal clusters.\n\nWhat's next\n-----------\n\n- Learn about [node taints to control scheduling.](/kubernetes-engine/docs/how-to/node-taints)\n- [Learn about Network tags](/vpc/docs/add-remove-network-tags).\n- [Learn about labels](https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/)"]]
