We are working with Apache Cassandra 3.11 and we are trying to prepare for a DR situation. For instance, we have two keyspaces that we back up via snapshot. For instance, via Ansible:
- name: Create snapshots
shell: "nodetool snapshot --tag {{ item.tag }} {{ item.name }}"
loop: "{{ keyspaces }}"
Now If we had to create a new cluster and use these snapshots, from what we understand, we would need to do the following:
- Create new cluster (3)
- Create keyspaces
- Restore each table via the snapshot
schema.cql - Use sstableloader
per node (only 2, third node has no snapshot), then
- Nodetool repair
per node
A reference to this path can be found here
However, the tables are loaded, but the data is duplicated, in our case, as we only have two nodes with snapshots as the third has none (failed node).
Are we following the correct path here? We have tried copying the snapshot data to the appropriate directory and used nodetool refresh -- keyspace table but the data never appeared.
Any ideas/help would be greatly appreciated.
