How to reduce pool size?

[hjlee9182]

Using the getCpunum function, the cpu of the node is all captured. The cpu of the node where our Milvus pops up is 128 cores, so the pool size is too large, so I want to reduce this.
But MiddlePriorityThreadCoreCoefficient can only be used as int. Shouldn't we let them use the floor value, too? (ex. 0.1)
Or I think the getCpuNum function should use only the cpu of that pod.

And is there any other way to adjust poolsize?

milvus/internal/querynodev2/segments/pool.go

Line 115 in 186a01e

poolSize := hardware.GetCPUNum() * pt.CommonCfg.MiddlePriorityThreadCoreCoefficient.GetAsInt()

[yhmo]

GetAsInt() returns an integer value, the minimum value is 1. It is illegal to set 0 or -1, or a float value.

[xiaofan-luan]

I'll fix this. We don't really need this pool anymore