(urgent) normality and classic assumption in var/vecm

[didik.pramusinto]

I have vecm model, unfortunately, I don't have enough knowledge on normality and classic assumption.
as long as I know, my model have problem in heteroscedasticity and autocorrelation, because the probability is lower than alpha that I use (5%). that's my first problem..
the second is: I'm using unit root test (ADF, level, none, lag=0) on the residual to justify that my mode is normal, is that process can be accepted?
please, any one give me suggestion on that... thanks before..

[didik.pramusinto]

There must be someone out there knowing the solution Please share...