Origin — Elementary Number Theory — Jones — p28 — Exercise 2.6
To instigate a contradiction, postulate $q_1,q_2,\dots,q_n$ as all the primes $\neq 2 (=$ the only even prime) of the form $3k+2$. Consider $N=3q_1q_2\dots q_n+2.$ None of the $q_i$ divides $N$, and $3 \not | N$.
$N$ = 3(the product of odd numbers) + 2 = odd number + even number = odd. Because $N \ge 2 $,
$\color{brown}{♯}$ because $N$ is odd $\implies 2 \not | N$,
thence by $\color{brown}{♯}$ and the Fundamental Theorem of Arithmetic, N = a product of one or more $\color{brown}{odd}$ primes.
The prime divisors of $N$ cannot be all of the shape $3k+1$. At least one of these primes is of the form $3k+2$ — Why? $(3a + 1)(3b + 1) = 3(...) + 1$, thence any number of (not necessarily distinct) primes of the form $3k+1$ is itself of the form $3k+1$.
But $N$ is not of the form $3k+1$. So some prime $p$ of the form $3k+2$ divides $N$. Overhead in first paragraph, we proved $q_i \not|N$ for all $i$. Therefore $p \notin$ $\{q_1,\dots,q_n\}$ of primes of the form $3k+2$, contradiction.
The general proof just starts with primes. Therefore how can you prefigure this proof's different start with odd primes ?
Where did this choice of $N$ hail from — feels uncanny?
I don't understand how none of the $q_i$ divides $N$?