have one node and connect to it with python sdk and running continues writes and reads.
When I have one node, Gets per second are up to 15 000, sets per seconds up to 5000. But when im adding new node and do balancing, Gets per second reduce to 10000 and sets per second to 2000.
What Im doing wrong please help.
Just to be sure, are you saying
What does the console show for each server for read/write activity?
Also keep in mind the latency of operations. If you’re running your python client on one of the nodes, the second node you add will have higher latency so if you don’t increase concurrency or work in larger batches, it’ll be normal for throughput to drop.
Another effect we’ve seen from time to time is different latencies on public cloud. For instance, the client could be more local to one node, but when provisioning another node it might be across AZ, etc.
Thanks for your reply.
No that reports are from couchbase Web monitoring tool.
So when Im reading with one node, just getting values by key, max OPS is 15- 16 k. If I add new node rebalance it and that do the same, READ OPS Decrease to 10k.
did you add replication ,after adding a node ?
No, Im just add node and rebalance it.
Where do you run your load generator?
What you mean under load generator?
There is something that is generating the workload where you’re measuring the performance. Where is that “load generator” running?