Using deprecated address after prefix removal from netdata about openthread HOT 2 CLOSED

Some ideas on how we may address the situation from the last comment (related to Domain ID use):

1. We can relax the check for domain ID in case the source address is not preferred (but the question become how to select if there happens to be multiple Domain IDs each with their own set of off-mesh routes published in netdata)?

2. We can check netdata entries and only if all entries belong to the same domain ID (which is the common/main case today), then allow the off-mesh route prefixes to be applied/used to determine next hop for the non-preferred source address (that does not match any prefixes)

3. We can add new mechanism to remember the domain ID the deprecating prefix/address belonged to and use the routes in netdata associated with the same ID (this seems like the best option to me).

4. Change the model and have the deprecating prefix still published in netdata (with its original domain ID). The downside is prefixes may end up piling up in the netdata (I am not personally in in favor of this approach).

From SPEC-1222 disucssion, we are leaning toward option 3.

from openthread.

Resolved by #9585

from openthread.