I'd be surprised if that were the issue, but of course I'd have to see the model. In all 3-D analysis software I've used, even if it doesn't simulate buckling affects on the overall model, it'll still flag the rod as "failed" when it compares the stress to the calculated allowable - which in this case would be essentially zero.
Now if the user thought "well, of course it fails in compression!" and ignored it, then yes...that could be a big problem. That would be so many levels of big problems. None of them surprising, sadly, but still big problems...