We’re planning to upgrade from Imply 2.6 to Imply 2.9 (Druid from 0.12 to 0.14) at AWS.
What’s the suggested approach. Rolling upgrade on existing cluster or Setting up new cluster and switch?
It seems setting up new cluster and switch from the existing to the new is more controllable and safer
though it takes time to loading segments for historicals at new cluster at start.
Like to collect your experience and opinions.