The rational roots theorem gives $\pm(1,2)$ as candidate roots. None of these actually produce zeros. Your recourse is to either rely on the cubic formula or a numerical method, as qt commented.
If you want to explore a numerical solution by hand, you can hone in on the solution with a bit of analysis.
\begin{align}
y&=t^3+2t^2+t-2
\\
\\
y'&=3t^2+4t+1
\\
&= (t+1)(3t+1)
\end{align}
The function has critical points at $t=-1,-\frac{1}{3}$. We can determine that both $y(-1)<0$ and $y(-1/3)<0$. Given that this is a cubic polynomial, we conclude that $y<0$ (i.e. no zeros) for $t<0$.
Since $y(0)=-2$ and $y'>0$ for $t>0$, we know that there is a single zero for $t>0$. By inspection, it must lie within $(0,1)$. You could try Newton's method from here, if you want to churn out a numerical solution by hand.