[Solr Wiki] Update of "NewSolrCloudDesign" by YonikSeeley

[Apache Wiki <wi...@apache.org>]

Dear Wiki user,

You have subscribed to a wiki page or wiki category on "Solr Wiki" for change notification.

The "NewSolrCloudDesign" page has been changed by YonikSeeley:
http://wiki.apache.org/solr/NewSolrCloudDesign?action=diff&rev1=9&rev2=10

Comment:
shard naming

  By deriving the top 8 bits of the hashcode from the user_id, it guarantees that any users emails are in the same 256th portion of the cluster.  At search time, this information can be used to only query that portion of the cluster.
  
  We probably also want to view the hash space as a ring (as is done with consistent hashing) in order to express ranges that wrap (cross from the maximum value to the minimum value).
+ 
+ === partition naming ===
+ When partitioning is by hash code, rather than maintaining a separate mapping of shard name (which is currently the actual core name), the shard name could actually be the actual range of documents within it.  So localhost:8983/solr/1-1000 would contain docs with a hash between 1 and 1000.  This is only necessary for auto-sharding using a hash on the id.  Other custom mappings should be able to take an id and map it directly to a partition.
  
  == Shard Assignment ==
  

Re: [Solr Wiki] Update of "NewSolrCloudDesign" by YonikSeeley

[Yonik Seeley <yo...@lucidimagination.com>]

I'm wondering if we shouldn't ditch the new term "partition" here and
just use "replica"?

In the past, we've sort of used "shard" to mean both a single physical
index, and the logical piece of the larger collection.  In practice,
this ambiguity normally isn't much of a problem as it's normally clear
by context and when it's not we sometimes throw in the word "replica".
 Examples:  "Doc X belongs on Shard Z", "Shard Z on this node is
corrupt".

Refreshing my memory on our ZK layout, it seems like we are using
"shards" in the logical sense there.

     /COLLECTIONS (v=6 children=1)
          COLLECTION1 (v=0 children=1) "configName=myconf"
               SHARDS (v=0 children=1)
                    SHARD1 (v=0 children=1)
                         ROGUE.LOCAL:8983_SOLR_ (v=0)
"node_name=Rogue.local:8983_solr url=http://Rogue.local:8983/solr/"

So perhaps we should just continue that, and change "partition" to
"replica" when necessary to prevent ambiguity?

-Yonik
http://www.lucene-eurocon.com - The Lucene/Solr User Conference

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org