One of my clients wants to have local nodes of the index. Just in case our
server go down.
Let me explain: say I have an index, call it "documents", spread of 20
shards, 5 nodes.
I want 2 of the nodes to be local to my client, and the remaining 3 local
to me. But.
I want their node to only contain their documents, and the remaining 3 to
contain all the other clients.
Is that possible without having to push the documents on 2 different
clusters (2 + 3)?
I thought about explicit shard routing, but I'm not sure how to
specifically route on the shard
(as I believe routing is done via hashing then modding on the complete
keyspace, which would
be "random", but also guarantees no hotspots on the cluster).
One of my clients wants to have local nodes of the index. Just in case our
server go down.
Let me explain: say I have an index, call it "documents", spread of 20
shards, 5 nodes.
I want 2 of the nodes to be local to my client, and the remaining 3 local
to me. But.
I want their node to only contain their documents, and the remaining 3 to
contain all the other clients.
Is that possible without having to push the documents on 2 different
clusters (2 + 3)?
I thought about explicit shard routing, but I'm not sure how to
specifically route on the shard
(as I believe routing is done via hashing then modding on the complete
keyspace, which would
be "random", but also guarantees no hotspots on the cluster).
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.