Hi,

I know these parameters can be set to tune the number of parallel scans on each region:

targetConcurrency, maxConcurrency, threadPoolSize, queueSize and maxIntraRegionParallelizaioin.

But I am wondering if there are some rules of thumb for setting their values optimally? Perhaps my cluster configuration is useful for you:

3 region servers, 8 cores and 8GB RAM on each. My table has 6 regions, 2 regions/region server, 5 million rows.

Thanks!

Reply via email to