Is it ever possible to make A and b equal to one in other oscillators
like RC-phase-shift oscillator etc.
Not with linear networks. This is like balancing a knife on its sharp edge - it falls one way (not oscillating) or it falls the other way (oscillating with AC distortion caused by gain stage overload).
Many excess-gain oscillators allow the amplifier non-linearity to set oscillator output amplitude. This is a simple way to ensure that oscillations start, and that a stable output amplitude is achieved. But output cannot be sinusoidal - harmonics are added because output amplitude wants to be higher than the linear amplifier can manage - sine peak(s) are compressed or clipped.
Do you want a clean, single-frequency sinusoidal output? Then automatic gain control should be added to the main gain stage. The main gain stage must have slightly excess gain to start oscillations. Without AGC, amplitude grows too much. AGC dials-back loop gain only at high amplitude to achieve that knife-edge balance.
Many textbooks quote that the product A*b must be always slightly
greater than one. wouldn't this lead to instability in the long run of
the circuit?
In OP's simple circuit, the gain stage is assumed perfectly linear. If you simulate this scenario in a circuit simulator, with a perfectly linear gain stage, then oscillation amplitude grows exponentially - I would call that "unstable". In practice, no amplifier is perfectly linear. Amplitude grows exponentially for a short time, but reaches some limit, defined by how the amplifier is powered.
If you dial back amplifier gain to exactly one, your circuit simulator might be satisfied to not oscillate. But if you give it a transient "kick" it might oscillate happily, with a very clean sinusoidal amplitude, although amplitude may slowly vary.
Unless this gain-of-one amplifier has a Thevenin output impedance of zero, you could not add a load that tries to extract power - it would cause loop gain to fall below one, and amplitude dies away exponentially. Again, unstable.
