I have a single 6-node datacenter right now, my main keyspace running replication_strategy Solr:6 and DSE 3.0.1.
I'd like to split those to better balance my workloads and I'm stuck on the actual procedure -- to start I want to move to Cassandra:3, Solr:3 and then scale each from there.
http://www.datastax.com/docs/datastax_enterprise2.2/deploy/workload_reprovisioning (DSE 3.0 link is broken)
States that solr nodes can't be reprovisioned which seems to be accurate as when I started up dse without the -s flag solr was turned on anyway.
My ring currently shows:
n0 Solr rack1 Up Normal 47.8 GB 16.67% 0
n1 Solr rack1 Up Normal 75.47 GB 16.67% 28356863910078203714492389662765613056
n2 Solr rack1 Up Normal 68.9 GB 16.67% 56713727820156407428984779325531226112
n3 Solr rack1 Up Normal 68.36 GB 16.67% 85070591730234615865843651857942052864
n4 Solr rack1 Up Normal 3.85 GB 16.67% 113427455640312814857969558651062452224
n5 Solr rack1 Up Normal 4.21 GB 16.67% 141784319550391032739561396922763706368
I thought I could essentially:
1. update replication stragegy
2. turn off solr on 3 nodes
3. generate new tokens for two datacenters
4. use nodetool move to move each node around the ring
5. repair each node
6. clean each node
Since #2 didn't work, what's the actual sequence?
P.S. any way to alter my profile nickname to not use my email address so it doesn't flow into the text of my posting?