In a NebulaGraph cluster environment, if the disk of a certain storage node becomes full, the exchange ingestion program will be unable to write data.
- Nebula version: v3.5.0
- Deployment mode: Distributed
- Installation method: Docker
- In production environment: Yes
- Hardware information:
- Disk: SATA
- Processor: 64 cores
- Memory: 128GB
There are 4 nodes in the storage cluster, and one of the storage nodes has run out of disk space. When submitting a spark-submit command to consume topic data in a spark-streaming program, the data cannot be stored due to the full data disk. How should this be handled to ensure that data can still be inserted even if a node's data disk is full?
