- Nebula version: 3.5
- Deployment mode: Distributed
- Installation method: RPM
- Production environment: No
- Hardware information:
- Disk: HDD
- CPU, Memory: 16G
- Description of the issue:
- I have deployed a three-node cluster (all with 16G memory) consisting of one meta node and three storage/graph nodes. The cluster has been load balanced. During the test query speed using the following statement:
match (v:productStock)-[e]-(v2)-[e2]-(v3) where v.pdstName=="中国银行" return count(v3)
Although there are only 5,000 nodes, it causes the graph service on the host to crash due to insufficient memory: Memory usage has hit the high watermark of the system, available: 1.40599e+08 vs. total: 16629743616 in bytes.
The other two nodes are functioning normally without any errors. Does this mean that the distributed setup is not working as expected?
Additionally, I have deployed the dashboard (using an older version, 0.17, of node_exporter) as per the official instructions. I can see the services, but I cannot see the machines. What could be the reason for this? enter image description here enter image description here
