The problem is that most textbooks oversimplify this topic to the point of being flat out wrong or very confusing.
Estrogens (that's a plural word) normally have negative feedback effects on LH and FSH during most of the estrous cycle (most non-human mammals) and the menstrual cycle (humans and a few non-human primates). So do progestins. However, LH and FSH (along with GnRH that is the hypothalamic neurohormone that regulates LH and FSH) are not just continually secreted. They are secreted in pulses (think "bursts" of secretory activity). And, inhibition by the sex steroids doesn't just suppress overall secretion, but actually alters the pulse pattern. Estrogens suppress LH pulse amplitude, but don't alter frequency (we won't discuss seasonally breeding animals here, because then even this statement needs qualifiers).. Progestins suppress pulse frequency. FSH pulses are a little messier to deal with.
The pattern is what determines how the ovary responds. In the predominantly estrogenic environment of the follicular phase, fast, low amplitude FSH and LH pulses tell the follicles to keep growing. When they get closer to ovulation, there is a short term switch to a positive feedback mechanism. I can't answer why...that's the million dollar question in research in reproductive endocrinology. We know a lot about ways to make it happen, but not fundamentally why. Anyway, that switch happens as the one dominant follicle gets very large and secretes heaploads of estrogens, and that in turn increases instead of decreases GnRH, LH and FSH secretion to the point where we can no longer detect individual pulses, but one big, massive, sustained secretory event that we call a surge. This triggers ovulation. Ovulation is the limiting step that keeps this positive feedback loop from getting out of control...ruptured follicle and no more follicular fluid producing estrogen. The remaining follicle cells undergo a conversion process and become the corpus luteum, which produces progesterone and some estrogens. Now LH and FSH are secreted in very slow pulses during the luteal phase, which doesn't stimulate any more follicle growth and instead allows the corpus luteum to function and produce more progesterone, which helps prep the uterus for implantation if pregnancy occurs. Not all species have luteal phases. Many rodents, for example, only form a functioning corpus luteum if pregnancy occurs and quickly prepare to ovulate again if they don't get pregnant (they have only 4-5 day estrous cycles).