“Success, but fitting stopped because change in residuals less than tolerance ” ？

Hi there,

I am trying to fit a set of data with Lorentzian by using fittype(). It turns out that only around 10% of the whole data are fitted well with a typical "output message" as

output = 

             numobs: 51
           numparam: 4
          residuals: [51x1 double]
           Jacobian: [51x4 double]
           exitflag: 1
      firstorderopt: 9.9194e-008
         iterations: 3
          funcCount: 20
       cgiterations: 0
          algorithm: 'trust-region-reflective'
            message: 'Success. Fitting converged to a solution.'

while for most of the data, the output is

output = 

             numobs: 51
           numparam: 4
          residuals: [51x1 double]
           Jacobian: [51x4 double]
           exitflag: 3
      firstorderopt: 2.3938e-006
         iterations: 3
          funcCount: 20
       cgiterations: 0
          algorithm: 'trust-region-reflective'
            message: 'Success, but fitting stopped because change in residuals less than tolerance (TolFun).'

Can anyone tell me the difference between 'Success. Fitting converged to a solution.' and 'Success, but fitting stopped because change in residuals less than tolerance (TolFun).' ?