AKS节点消耗 - 超时/无响应 [英] AKS Node drain - timing out/no response
问题描述
我正在尝试通过驱逐节点(封锁/排放/删除)来升级失败的AKS集群的节点。
I'm trying to upgrade the nodes of a failed AKS cluster, by evicting nodes (cordon/drain/delete).
步骤是:
- 将AKS群集升级到所需版本
- Cordon a Node(kubectl cordon)
- 排空节点(kubectl排水)
- 删除节点(kubectl删除)
- 扩展新的新版本的节点(az aks scale)
我已经成功地以这种方式升级了几个集群。 对于一个特定的群集,我不得不耗尽第一个节点。 drain命令永远不会返回提示符(这是来自Cloud Shell)。 Cloud Shell最终会超时。
I've been successful in upgrading a few clusters this way. For one particular cluster, I'm stuck on draining the first node. The drain command never returns a prompt (this is from Cloud Shell). And Cloud Shell will eventually time out.
有关如何继续的建议吗?
Any suggestions on how to proceed?
推荐答案
从过去的问题我看到,如果你无法让群集摆脱失败的状态,解决方案通常是删除群集并重新部署。
From past issues I have seen that if you cannot get a cluster out of a failed state the resolution is generally to delete the cluster and redeploy it.
您是否尝试将--force命令添加到排水管以查看是否有帮助?
Have you tried adding the --force command to the drain to see if it helps?
这篇关于AKS节点消耗 - 超时/无响应的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!