...
- Version: based onFlink 1.19
- Deployment mode: yarn per-Job
- Flink yarn cluster:
- 1 master 2 workers
- specifications:
- master: ecs.g7.2xlarge 8 vCPU 32 GiB (Alibaba Cloud)
- worker: ecs.i2g.2xlarge 8 vCPU 32 GiB (Alibaba Cloud)
- HDFS cluster:
- 1 master 2 workers
- specifications:
- master: ecs.g7.2xlarge 8 vCPU 32 GiB (Alibaba Cloud)
- worker: ecs.i2g.2xlarge 8 vCPU 32 GiB (Alibaba Cloud)
- yarn cluster and HDFS cluster are on the same LAN
- State backend: ForSt (Based on RocksDB 8.5.3)
- Job config:
- Memory: Task Manager 4GB3GB, Job Manager 4GB1600MB
- Checkpointcheckpoint: Disabled (Not fully supported in PoC yet)
...