I have a setup with a “large” VM running as the console with 16gig RAM and 8 cores, and 10x smaller VMs running as agents. I’m trying to build very large DBs - 100GB, 500GB, 1TB, 5TB etc. The TPC-C load time is considerable - days. This is all to test DB performance with our self-designed storage system - however as I build the DB, the IOPS we’re getting on our system are very low - in the hundreds of IOPS. We can easily deliver more than that. I reduced the thinking time option, and turned on the distributed index option.
Am I missing anything? What is the recommended setup of console and agents? How many?
I’m seeing similar behavior with SQL Server, MySQL, Oracle…