Multicore, shards, merging indexes, SolrCloud: Which to use?
I'm hoping that someone can tell me what is the best way to handle the
following scenario:
I have three types of entities (e1, e2, e3) in a database, each in their
own table. These entities have their own primary key (e1.[k1], e2.[k2],
e3.[k3]) and some common fields and some unique ones.
e1.[k1], e1.k2?, e1.k3?, e1.commonfield, e1.uniquefield
e2.[k2], e2.k3?, e2.commonfield
e3.[k3], e3.commonfield
Where: [primary key], optional?
Currently, I'm using four cores, three of them map onto these entities,
and have their own import/deltaImport that uses the primary key of that
specific entity to do the import. The fourth core uses the other three as
shards, so that all three entity types can be searched at the same time.
This requires the schema for all four cores to be identical.
[id], k1?, k2?, k3?, commonfield, uniquefield?
The Solr document id is generated using CloneFieldUpdateProcessorFactory /
ConcatFieldUpdateProcessorFactory by concatenating the primary key values.
I'm doubtful if this is the correct way to go about it. Technically, this
is not a candidate for sharding, but I'm unsure what other approach I
could or should follow. For this particular example, is this a good
approach? Are there better options?
Thank you kindly.