There is a bit of a gap in your understanding. Zookeeper as such doesn’t hold any data. Zookeeper holds only the Solr Collections configs and cluster state/node status. In short, zookeeper lets every other Solr node know whether other replica nodes are up and running.
What process takes place when requesting information within Solr?
- If you are using Solr client, your client will eventually fetch data from one of the up/running solr nodes among the replicas.
Further, the “DB” you are talking about is a bunch of index files that are stored in solr nodes and not zookeeper.
Solr client talks to only Solr nodes for any request, read/write. Solr client talks to zookeeper to check which solr nodes are up and running and not to fetch data.