I have a server with DruidDB release 0.15.1 with a datasource that contains 128 millions of records.
Now I want to upgrade to release 0.17.
Can I replace only Druid executables or have I reload datasource ?
Are there compatibility issues between 0.15 and 0.17 releases?
I use kafka ingestion to load datasource and perform query by SQL with JSON over HTTP.
There is one thing related to Select vs Scan queries where Select is deprecated in in druid-0.17.
You might have to rewrite your select queries with Scan.
You don’t need to reload data sources .
Of course , you will see some improvements related to index_parallel native batch ingestions, compactions, parallel merging feature related to group by queries on brokers etc