We are planning to shrink our cluster from 18 nodes to 12 nodes (much beefier).
Since i dont have too much free storage i dont have the luxury of running "rebalance_cluster()" against 100% of my data by marking 6 nodes ephemeral. So this is what i was planning to do:
- Manually (one at a time) copy large tables from 18 nodes to 12 nodes. This should make for 80% of the data.
- Set node#13-18 as ephemeral
- Kick off rebalance_cluster() on the other 20% of data.
When i kick off rebalance_cluster() would vertica know that it has to rebalance just the other 20% of data?
As such will these steps/process work in my situation?
Any pointers would be much appreciated
