I have setup a SolrCloud with 2 nodes, 2 shards and 2 replicas. My Join query works fine in a single solr instance and it has some issues with this SolrCloud setup.
fq = {!join from=student_id_i to=student_id_i} sf_meta_class:Students
q = sf_meta_class: Books
I guess mostly the issue is with the Sharding part, since I get nearly(not exactly) half of the result set and not the whole result set. Am I doing anything wrong here? Or the Solr never supported Join queries with multiple shards?
I had tried putting the shards explicitly on the search query like this:
Still it doesn't work for me.
EDIT: Apparently SolrCloud doesn't join queries in Shards. Found it on https://cwiki.apache.org/confluence/display/solr/SolrPerformanceProblems
Unfortunately Solr doesn't support this. An issue has been created, but in the meantime if you can reduce either set of docs to its own collection (for example separate sf_meta_class:Books) Solr does support joining across collections. The caveat there is you need to make sure that the smaller collection is not sharded and is present on each node.