SLSQP Optimization Early Termination due to - Positive directional derivative for linesearch

Question

I have an 40-Dimensional Rosenbrock function with nonlinear and bound constrains. I am using the following code to get minimum. I let internal finite-difference handle the gradient and I am using scaled variables. I do not understand why output gives me the "Positive directional derivative for linesearch" termination.

I played with initial guess but it did not work well. Since I am working with scaled variables optimization variables bounds must always lb=0 and ub=1. Following is my code.

import numpy as np
from scipy.optimize import minimize, NonlinearConstraint,Bounds

def Denormalizer(u_norm,maximum,minimum):

    u1 = u_norm*(maximum - minimum) + minimum

    return u1

def rosenbrock_40d(x):
    x=Denormalizer(x,maximum,minimum)
    """The Rosenbrock function"""
    return sum(100.0*(x[1:]-x[:-1]**2.0)**2.0 + (1-x[:-1])**2.0)

def nonlinear_constraint(x):
    x=Denormalizer(x,maximum,minimum)
    sum_of_squares = 0
    for i in range(39):
        sum_of_squares += 0.1-(x[i]-1)**3-(x[i+1]-1)
    return sum_of_squares

nlcfd = NonlinearConstraint(nonlinear_constraint, -np.inf,0)
Opt_Bounds = Bounds(0,1)
x_values = [
    2, -8, 5, 10, -3, 7, -9, 1, -6, 4,
    -2, 8, -7, 3, 6, -5, 9, -1, 0, -4,
    7, -3, 6, -10, 8, -2, 9, -5, 4, -1,
    0, -6, 10, -8, 3, -4, 2, -7, 1, -9
]
maximum,minimum=np.max(x_values),np.min(x_values)
x_values_norm = (x_values - np.min(x_values)) / (np.max(x_values) - np.min(x_values))
    
resultfd = minimize(rosenbrock_40d, x_values_norm, method='SLSQP',bounds=Opt_Bounds, constraints=nlcfd,options={'maxiter':200,'disp':True,'iprint':2,'ftol':1e-6})

This is the output.

  NIT    FC           OBJFUN            GNORM
    5    41     8.636645E+06     2.061750E+07
Positive directional derivative for linesearch    (Exit mode 8)
            Current function value: 8636645.0
            Iterations: 5
            Function evaluations: 41
            Gradient evaluations: 1