Auf dieser Seite erfahren Sie, wie Sie die Netzwerkbandbreite für GPU-Knoten in GKE-Clustern (Google Kubernetes Engine) mithilfe von Google Virtual NIC (gVNIC) erhöhen.
In Autopilot-Clustern wird auf Knoten, auf denen die GKE-Version 1.30.2-gke.1023000 oder höher ausgeführt wird, automatisch eine Google Virtual NIC (gVNIC) installiert.
Die Anleitung auf dieser Seite gilt nur für Standardcluster.
Hinweise
Führen Sie die folgenden Schritte durch, bevor Sie beginnen:
Wenn Sie die Google Cloud CLI für diese Aufgabe verwenden möchten, müssen Sie die gcloud CLI installieren und dann initialisieren. Wenn Sie die gcloud CLI bereits installiert haben, rufen Sie die neueste Version mit gcloud components update ab.
Die gVNIC API wird nur auf GPU-Knoten unterstützt.
Wenn Sie die Bandbreite auf Nicht-GPU-Knoten erhöhen möchten, aktivieren Sie Stufe-1-Bandbreite.
Voraussetzungen
GKE-Knoten müssen ein Knoten-Image für Container-Optimized OS verwenden.
gVNIC aktivieren
Sie können einen Cluster mit Knotenpools erstellen, die gVNIC verwenden, einen Knotenpool mit aktivierter gVNIC erstellen oder einen Knotenpool aktualisieren, um gVNIC zu verwenden.
Cluster erstellen
Erstellen Sie einen Cluster mit Knotenpools, die gVNIC verwenden:
[[["Leicht verständlich","easyToUnderstand","thumb-up"],["Mein Problem wurde gelöst","solvedMyProblem","thumb-up"],["Sonstiges","otherUp","thumb-up"]],[["Schwer verständlich","hardToUnderstand","thumb-down"],["Informationen oder Beispielcode falsch","incorrectInformationOrSampleCode","thumb-down"],["Benötigte Informationen/Beispiele nicht gefunden","missingTheInformationSamplesINeed","thumb-down"],["Problem mit der Übersetzung","translationIssue","thumb-down"],["Sonstiges","otherDown","thumb-down"]],["Zuletzt aktualisiert: 2024-11-21 (UTC)."],[],[],null,["# Increase network traffic speed for GPU nodes\n\n[Autopilot](/kubernetes-engine/docs/concepts/autopilot-overview) [Standard](/kubernetes-engine/docs/concepts/choose-cluster-mode)\n\n*** ** * ** ***\n\nThis page shows you how to increase network bandwidth for GPU nodes on\nGoogle Kubernetes Engine (GKE) clusters by using\n[Google Virtual NIC (gVNIC)](/compute/docs/networking/using-gvnic).\n\nIn Autopilot clusters, nodes that run GKE version 1.30.2-gke.1023000\nand later have Google Virtual NIC (gVNIC) installed automatically.\nThe instructions on this page only apply to Standard clusters.\n\nTo increase bandwidth on CPU nodes, consider enabling\n[Tier-1 bandwidth](/kubernetes-engine/docs/how-to/using-tier-1#enable_tier_1_bandwidth).\n\nBefore you begin\n----------------\n\nBefore you start, make sure that you have performed the following tasks:\n\n- Enable the Google Kubernetes Engine API.\n[Enable Google Kubernetes Engine API](https://console.cloud.google.com/flows/enableapi?apiid=container.googleapis.com)\n- If you want to use the Google Cloud CLI for this task, [install](/sdk/docs/install) and then [initialize](/sdk/docs/initializing) the gcloud CLI. If you previously installed the gcloud CLI, get the latest version by running `gcloud components update`. **Note:** For existing gcloud CLI installations, make sure to set the `compute/region` [property](/sdk/docs/properties#setting_properties). If you use primarily zonal clusters, set the `compute/zone` instead. By setting a default location, you can avoid errors in the gcloud CLI like the following: `One of [--zone, --region] must be supplied: Please specify location`. You might need to specify the location in certain commands if the location of your cluster differs from the default that you set.\n\n### Limitations\n\n- [Compute Engine](/compute/docs/networking/using-gvnic#limitations) limitations apply.\n\n### Requirements\n\n- GKE nodes must use a Container-Optimized OS [node image](/kubernetes-engine/docs/concepts/node-images#cos).\n\nEnable gVNIC\n------------\n\nYou can create a cluster that has node pools that use gVNIC, create a node pool\nwith gVNIC enabled, or update a node pool to use gVNIC.\n\n### Create a cluster\n\nCreate a cluster with node pools that use gVNIC: \n\n gcloud container clusters create \u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e \\\n --accelerator type=\u003cvar translate=\"no\"\u003eGPU_TYPE\u003c/var\u003e,count=\u003cvar translate=\"no\"\u003eAMOUNT\u003c/var\u003e \\\n --machine-type=\u003cvar translate=\"no\"\u003eMACHINE_TYPE\u003c/var\u003e \\\n --enable-gvnic\n\nReplace the following:\n\n- \u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e: the name of the new cluster.\n- \u003cvar translate=\"no\"\u003eGPU_TYPE\u003c/var\u003e: the type of [GPU accelerator](/compute/docs/gpus) that you use. For example, `nvidia-tesla-t4`.\n- \u003cvar translate=\"no\"\u003eAMOUNT\u003c/var\u003e: the number of GPUs to attach to nodes in the node pool.\n- \u003cvar translate=\"no\"\u003eMACHINE_TYPE\u003c/var\u003e: the type of machine you want to use. gVNIC is not supported on [memory-optimized machine types](/compute/docs/machine-types#memory-optimized_machine_type_family).\n\n### Create a node pool\n\nCreate a node pool that uses gVNIC: \n\n gcloud container node-pools create \u003cvar translate=\"no\"\u003eNODEPOOL_NAME\u003c/var\u003e \\\n --cluster=\u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e \\\n --enable-gvnic\n\nReplace the following:\n\n- \u003cvar translate=\"no\"\u003eNODEPOOL_NAME\u003c/var\u003e: the name of a new node pool.\n- \u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e: the name of the existing cluster.\n\n### Update a node pool\n\nUpdate a node pool to use gVNIC: \n\n gcloud container node-pools update \u003cvar translate=\"no\"\u003eNODEPOOL_NAME\u003c/var\u003e \\\n --cluster=\u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e \\\n --enable-gvnic\n\nReplace the following:\n\n- \u003cvar translate=\"no\"\u003eNODEPOOL_NAME\u003c/var\u003e: the name of the node pool that you want to update.\n- \u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e: the name of the existing cluster.\n\nThis change requires recreating the nodes, which can cause disruption to your\nrunning workloads. For details about this specific change, find the\ncorresponding row in the [manual changes that recreate the nodes using a node\nupgrade strategy without respecting maintenance\npolicies](/kubernetes-engine/docs/concepts/managing-clusters#manual-changes-strategy-but-no-respect-policies)\ntable. To learn more about node updates, see [Planning for node update\ndisruptions](/kubernetes-engine/docs/concepts/managing-clusters#plan-node-disruption).\n| **Caution:** GKE immediately begins recreating the nodes for this change using the node upgrade strategy, regardless of active maintenance policies. GKE depends on [resource\n| availability](/kubernetes-engine/docs/how-to/node-upgrades-quota) for the change. Disabling node auto-upgrades [doesn't prevent this\n| change](/kubernetes-engine/docs/how-to/node-auto-upgrades#disable-only-versions). Ensure that your workloads running on the nodes are prepared for disruption before you initiate this change.\n\nDisable gVNIC\n-------------\n\nUpdate the node pool using the `--no-enable-gvnic` flag: \n\n gcloud container node-pools update \u003cvar translate=\"no\"\u003eNODEPOOL_NAME\u003c/var\u003e \\\n --cluster=\u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e \\\n --no-enable-gvnic\n\nThis change requires recreating the nodes, which can cause disruption to your\nrunning workloads. For details about this specific change, find the\ncorresponding row in the [manual changes that recreate the nodes using a node\nupgrade strategy without respecting maintenance\npolicies](/kubernetes-engine/docs/concepts/managing-clusters#manual-changes-strategy-but-no-respect-policies)\ntable. To learn more about node updates, see [Planning for node update\ndisruptions](/kubernetes-engine/docs/concepts/managing-clusters#plan-node-disruption).\n| **Caution:** GKE immediately begins recreating the nodes for this change using the node upgrade strategy, regardless of active maintenance policies. GKE depends on [resource\n| availability](/kubernetes-engine/docs/how-to/node-upgrades-quota) for the change. Disabling node auto-upgrades [doesn't prevent this\n| change](/kubernetes-engine/docs/how-to/node-auto-upgrades#disable-only-versions). Ensure that your workloads running on the nodes are prepared for disruption before you initiate this change.\n\nTroubleshooting\n---------------\n\nTo troubleshoot gVNIC, see\n[Troubleshooting Google Virtual NIC](/compute/docs/troubleshooting/gvnic).\n\nWhat's next\n-----------\n\n- Use [network policy logging](/kubernetes-engine/docs/how-to/network-policy-logging) to record when connections to Pods are allowed or denied by your cluster's [network policies](https://kubernetes.io/docs/concepts/services-networking/network-policies/)."]]
