I have some questions on MongoDB setup model, and hope every one here can give me hints/advices.
Currently i’m running 3 bare metal for one primary - master and two secondary for replicaSet. However i want to deploy a shard model for the data to be sharded into two replicaSets, to have this setup i will put more one bare metal into three existing servers. So below is my hardware plan:
03 bare metal servers with 6 cores that use in current DB system:
- AMD 4.5GHZ Ryzen 7 (12 core virtually) with 32 GB RAM, 512 MVME 512GB 4 x PCIe3,
- these servers will be breakdown into 1 for mongos (router) server, and 02 for replicaSet server
01 new bare metal:
- 4 cores Intel i5 3,2 Ghz CPU, 16GB RAM, 256 GB SSD
- This server wil be configSrv only.
This is the minimum setup for lab testing
My question is: Will this new setup create any advantages over the old setup with one primary with two secondary i’m using now in term of performace? what is the advantages of having two mongos instead one just in case of redundancy only, any performace change if we have two mongos? Will the two mongos will really distribute the request from application to the two shards?
Data will be exported form old DB and imported into new system after setup, nothing change in data structure
I think two shards with new setup are enough for now, if need i can add more later.
I’m plannimg this new setup for deployment and development test before production
Thank you very much.