This is not reproducible, and you have not told us the seed you used which
is what is needed to make random results reproducible.
Please supply a reproducible example (if you can). I ran the first line
100 times (about 15 minutes) without encountering any NAs. If you found
these fairly easily it indicates a bug in your Windows installation, not
one in R. Further, if it were at all common then other people would have
encountered it and surely mentioned it.