[Freeipa-devel] DNA range distribution to replicas by default

[Simo Sorce]

----- Original Message -----
> From: "thierry bordaz" <tbordaz at redhat.com>

> About the ranges, each replica has a unique replicaID, the selection of
> the ranges could use this replicaID for most significant digit.

This doesn't really work unless you know before hand how many replicas you
have to partition the space.

> Publishing the ranges to the shared tree looks good but what is benefit
> of publishing dnaRemainingValues (either the exact value or sample) ?
> Who is consuming it ?

1. Servers that have their range depleted use that to figure out what other
sever they can contact to get a new slice.
2. Admins that want to monitor range consumption at a coarse level to plan
whether they need to add whole new ranges to the domain.

Simo.