I all, we are prepping for our first prod deployment, and are setting up our meta-datadb in mysql.
I do not see anything regarding replication setup at: https://druid.apache.org/docs/latest/development/extensions-core/mysql.html
We will obviously have 2 mysql databases, my question is, should we set these up as master/master? can druid handle that? or do we need to master<-slave replication?
also, side question…which druid processes actually talk to the meta-db?
Master slave replication definitely works. Haven’t tried the master-master set up and not sure if it works.
which druid processes actually talk to the meta-db?
From the docs https://druid.apache.org/docs/latest/dependencies/metadata-storage.html,
The Metadata Storage is accessed only by:
Indexing Service Processes (if any)
Realtime Processes (if any)
Thanks for the info, I’m a bit concerned with the 1 point of failure built into the system, but I suppose that if I ever get the time, I could contribute to the project, by improving that situation.