Does kubernetes deployment "request memory" consume real resources

I have one question about resources usage and request:

Let us say I have only one k8s node. It has 12GB memory, 2GB is used by system components. The other 10GB are using by k8s pods.
Assume I already have two pods on this nodes
1 First pod memory request is 7GB, real memory usage is only 1GB
2 Second pod memory request is 1MB, real memory usage is 3GB.

My question are:

1 Will the node raise memory low usage issue?

2 Whether “pod ==> resource => request => memory” are consuming real memory or not?

3 Now, if I create a new pod whose request memory is 100MB, can it be scheduled on this node?