Hi Dulce, MiniCluster is generally used for local testing and is limited by the resources of a single machine. When more tasks are executed, it may not be able to immediately acquire the resources needed to start the MiniCluster, resulting in slower startup times.
If running Flink tasks in a production environment, it is recommended to use cluster deployment mode[1]. You can also use resource providers like Yarn[2] or Kubernetes[3]. Best, Jiabao [1] https://nightlies.apache.org/flink/flink-docs-release-1.18/docs/deployment/resource-providers/standalone/overview/ [2] https://nightlies.apache.org/flink/flink-docs-release-1.18/docs/deployment/resource-providers/yarn/ [3] https://nightlies.apache.org/flink/flink-kubernetes-operator-docs-main/ On 2024/01/17 21:11:56 Dulce Morim wrote: > Hello, > > In a single JVM, I'm running multiple flink batch jobs locally using the > MiniCluster (Java 17 and Flink 1.18). > > At the beginning of the process, the Mini Cluster starts pretty much > instantly. However, I'm running into an issue where the more jobs I execute > the longer the MiniCluster takes to start. > > Here's an example: > > 2024-01-17 17:07:26.989 [INFO ] MiniCluster - Starting Flink Mini Cluster > 2024-01-17 17:07:27.165 [INFO ] MiniCluster - Starting Metrics Registry > 2024-01-17 17:07:33.801 [INFO ] MetricRegistryImpl - No metrics reporter > configured, no metrics will be exposed/reported. > 2024-01-17 17:07:33.801 [INFO ] MiniCluster - Starting RPC Service(s) > 2024-01-17 17:07:34.646 [INFO ] MiniCluster - Flink Mini Cluster started > successfully > > Has anyone faced this issue? > > Thanks. >
