Percona cluster CPU usage


I have 2 nodes running with the following specifications:

  • 10 CPUs
  • 16GB RAM
  • 8GB Swap

The “Threads_connected” count is approximately 80, and the CPU values of the hosts are insufficient. I notice that certain long-running queries are causing freezes at regular intervals. Do you have any suggestions other than adding CPU resources?

Threads_connected doesn’t matter; that’s just the total number of connections. What really matters is Threads_Running; these are threads actually doing work. You can have 10,000 T_connected, and only 10 T_running.

Please setup so we can see a much better picture of your CPU/Memory load.

No database server should have swap. If you are swapping to disk, then you need more RAM.

1 Like