I have the following matrix:
$$ \begin{pmatrix} 3&0&0\\ k+2&1&0\\ 5&k-1&1 \end{pmatrix} $$
The exercise asks to find the eigenvalues of the matrix and, for all $k\in\mathbb{R}$, determine a basis of each eigenspace of the matrix.
Since ths is a lower triangular matrix, the eigenvalues are the values on the diagonals: $\lambda=3$ and $\lambda=1$.
Now, for $\lambda=3$ I need to find the null space of this matrix:
$$ \begin{pmatrix} 0&0&0\\ k+2&-2&0\\ 5&k-1&-2 \end{pmatrix} \rightarrow \begin{pmatrix} 5&k-1&-2\\ k+2&-2&0\\ 0&0&0 \end{pmatrix} $$
By reducing in row echelon form ($R_2\leftarrow5R_2-kR_1$ and then $R_2\leftarrow R_2-2R_1$):
$$ \begin{pmatrix} 5&k-1&-2\\ 0&-k^2-k-8&2k+4\\ 0&0&0 \end{pmatrix} $$
I'm stuck here because $-k^2-k-8$ doesn't have real roots. Any hints?
Since $-k^2-k-8$ can’t be zero, you can safely divide by it and continue merrily on your way with the row-reduction. However, because you’re working in $\mathbb R^3$ there’s a much simpler way to find a basis for this null space. The row space of $A-3I$ is clearly spanned by its two nonzero rows, and the null space of a matrix is the orthogonal complement of its row space, so the null space of $A-3I$ is spanned by $$(k+2,-2,0)\times(5,k-1,-2) = (4,2k+4,k^2+k+8).$$