如何扩大/缩小正在运行的 Flink 集群? [英] How do I scale up/down a running Flink cluster?
问题描述
我现在在 Kubernetes 上运行 Flink.我假设如果我更新 TaskManager 部署的副本，Kubernetes 会为我增加/减少 TM pod 的数量，但我不确定这是否就是我需要做的全部.我是否需要做任何其他事情来使工作适应更多/更少的 TM?
I'm running Flink on Kubernetes now. I assume if I update the replicas of TaskManager deployment, Kubernetes scales up/down the number of TM pods for me, but I'm not sure if that is all I need to do. Do I need to do anything else to make the job adapt to the more/less TMs?
推荐答案
默认情况下，Apache Flink 不会根据任务管理器数量的变化重新缩放.
Apache Flink does not, by default, rescale in response to changes in the number of task managers.
关于 Flink 如何在 K8s 环境中重新缩放有多种方案.一种称为主动模式"，是 Flink 知道它想要什么资源，并与 K8s 配合以相应地获取/释放资源.请参阅如何在本地部署 FlinkKubernetes 了解详情.
There are various schemes for how Flink rescales in a K8s environment. One, referred to as "active mode", is where Flink knows what resources it wants, and works with K8s to obtain/release resources accordingly. See how to deploy Flink natively on Kubernetes for details.
另一种方案，称为反应模式"，正是您所希望的:Flink 扩展自身以使用任何可用的资源.有关详细信息，请参阅弹性缩放.
Another scheme, referred to as "reactive mode", is what you appear to be hoping for: Flink scales itself to use whatever resources have been made available. See Elastic Scaling for details.
您还可以在 Kubernetes 上部署 Flink，这实际上是独立部署的容器化版本.在这样的环境中，重新缩放涉及以下步骤:
You can also deploy Flink on Kubernetes in what is effectively a containerized version of a standalone deployment. In such an environment, rescaling involves these steps:
- 在获取保存点时停止作业
- 从保存点恢复作业，并为新集群安排了适当的大小
Flink Operations Playground 有一个分步示例，说明如何在基于 docker-compose 的部署中进行重新缩放.
The Flink Operations Playground has a step-by-step example of how to do rescaling in a docker-compose-based deployment.
这篇关于如何扩大/缩小正在运行的 Flink 集群?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！
