>2**127 is correct (since MD5 is a 128-bit hash).

Ya, I had just thought that the initial token was interpreted by Cassandra's interpreter as the exponent, not as a literal value.

Thanks,
Robin.

-----Original Message-----
From: Jonathan Ellis [mailto:jbellis@gmail.com] 
Sent: December 8, 2009 8:28 PM
To: cassandra-user@incubator.apache.org
Subject: Re: Configuring Cassandra

On Tue, Dec 8, 2009 at 7:05 PM, Coe, Robin <robin.coe@bluecoat.com> wrote:
> I'd love to see this document grow as time permits, to include known
> caveats, hidden features, tweaks, etc.  Perhaps, with some common use
> cases to help elucidate the content?

My writing well is exhausted for now but that's the nice thing about wikis... :)

> One question, though.  I thought I read about using an initial token for
> a RandmizedPartioner of between 0 and 127, but in your write-up, you say
> the value should be i * (2**127 / N) for i = 1 .. N.  That's a slightly
> larger number.  Can you clarify?

2**127 is correct (since MD5 is a 128-bit hash).