Article says:“IO Threads were at 2 writer and 1 reader thread”. I would like to adjust various number of writers vs reader thread and see how it affect the cluster, but I have no idea how to do this.
Do you know what specifically I need to do to change this setting?
Article says:“Disk queue capacity was set to 5M items”. Same thing here, how do I adjust this value?
Hi @steevebisson, these are on github - look for “increase disk cap”.
Looks like we missed the reader/writer config we used in the script.
One thing I’ll mention is, we pushed this cluster to its limits. and typically production workloads would leave some room in capacity to absorb variations or to allow operations like rebalance to run without impacting the incoming workload greatly. The configurations were done for us to squeeze the last drop for the case we are running which is a pure write workload. In a production system, you won’t need these settings. We tune our defaults to be able to run mixed workloads and concurrent operations with read/writes and queries, rebalance operations and more. So defaults will get you a better balance on performance.
thanks
-cihan