Currently not in plan - just looking to provide the ability for other
nodes to take over if one node drops dead (and to do so by first
crash-recovering the in-memory component of that node - so we don't plan
have multiple queryable replicas - just one - we'll load balance by
partitioning and not by replicating).
On 11/2/15 9:46 AM, Wail Alkowaileet wrote:
Dears,
I know Murtadha is working on introducing replication on Asterix. I want to
ask if that feature will also affect the Result Distribution Framework ? in
other words, are we going to have a replicated result-set ?
