Yes.
The validation set is just being used how well the trained model works on examples it hasn’t seen during training, and so it being shuffled is irrelevant.
Yes.
The validation set is just being used how well the trained model works on examples it hasn’t seen during training, and so it being shuffled is irrelevant.