On Tue, 2011-06-14 at 09:48 -0400, Simo Sorce wrote:
> On Tue, 2011-06-14 at 07:42 -0400, Stephen Gallagher wrote:
> > The decision to make the range start at 1 billion was made
> > specifically
> > BECAUSE the chances of a company having that many users was
> > statistically unlikely.
> Correction we start at 1Million and we get a 100k range randomly within
> the 1M-2B range, so almost 10k different possible buckets.
Ah I must correct myself, I changed the values before the 2.0 release so
the random range is 200k-2B which makes up the 10k buckets :-)
The code is actually this:
namespace = random.randint(1, 10000) * 200000
> The chance 2 installations end up getting the same bucket are very low.
> owever you can always force the UID to be used at user creation by
> explicitly specifying the IDs you want.
Simo Sorce * Red Hat, Inc * New York
Freeipa-users mailing list