The thrust loads across a fan disk are insignificant to the loads caused by spinning.
Vibratory stresses are typically way below any measurable contribution to the tangential stresses in a fan disk.
The materials used in the fan disk and the design stresses are very resistant to crack growth in a few cycles.
Whatever happened is likely way outside simple assumptions.
When you design a bolt flange attachment in a high stress area you have to avoid concentrating stress fields around only a few bolt holes. You do this by adding extra holes (typically not used) to smooth out the stress field thus avoiding fatigue fatigue cracking from these holes.