0
votes

I have one cassandra datacenter. Let's name it DC1. Then I added new datacenter for extending purporse in nodesize. Let's name it DC2. I use replication_factors DC1:3 and DC2:3. I write all my data as LocalDC=DC2 and ConsistencyLevel.LocalQuorom. I am sure that all write requests go to DC2. I want to remove DC1 and I dont want to run nodetool repair command. I dont want to wait.

Can I just simply change all keypaces replication_factor DC2:3 and run nodetool decommission on DC1 nodes?

1

1 Answers

0
votes

Yes.

As you said you are sure that there is no data latency between two data centers you can skip that step.

  1. Change all key space replication strategy using ALTER
  2. Decommission all the nodes one by one

See this: https://docs.datastax.com/en/cassandra/2.0/cassandra/operations/ops_decomission_dc_t.html