A
Andriy Mishin
Guest
Hello, It was a very interesting discussion. Thank you! About the MTX latch I also see large values per second for this database. That's not good? About 128 CPU. There are 15 databases running on this server. These fifteen databases are part of one large application. And it's only in one branch. A separate server is used for each branch. My colleagues told me that earlier, a long time ago, it was a single database. But because of the very high MTX, they had to eventually split this single database into several in order to spread the load and resource utilization, including for MTX. Also, each client connects to all fifteen databases at once when logging in to the application. For this database is set -Mn=900/-Mi=1/-Ma=30. However, I don't know why, for other databases on this server the value of -Ma is 1. That is, in this database one server process serves up to 30 clients, while in other databases one server process serves only one client. Is it possible that because of -Ma=30 we have high MTX and BHT latches in this database? Also, if I understand correctly, one physical server is running at the same time -Mn=900 * 15 DB server processes for remote clients. Wouldn't it be better if we had a lot of CPU for better performance?
Continue reading...
Continue reading...