Druid configuration files directory structure

Hi,

I am confused on the configuration files directory structure. The recommeded configuration file organization page talks about having conf directory structure as below:

conf/druid:
_common       broker        coordinator   historical    middleManager overlord

conf/druid/_common:
common.runtime.properties log4j2.xml

conf/druid/broker:
jvm.config         runtime.properties

conf/druid/coordinator:
jvm.config         runtime.properties

conf/druid/historical:
jvm.config         runtime.properties

conf/druid/middleManager:
jvm.config         runtime.properties

conf/druid/overlord:
jvm.config         runtime.properties

conf/tranquility:
kafka.json  server.json

But when I see the Druid 0.15.0 version release package, I see below strcuture:

conf/druid:
cluster single-server

conf/druid/cluster:
_common data master query

conf/druid/cluster/_common:
common.runtime.properties log4j2.xml

conf/druid/cluster/data:
historical middleManager

conf/druid/cluster/data/historical:
jvm.config main.config runtime.properties

conf/druid/cluster/data/middleManager:
jvm.config main.config runtime.properties

conf/druid/cluster/master:
coordinator-overlord

conf/druid/cluster/master/coordinator-overlord:
jvm.config main.config runtime.properties

conf/druid/cluster/query:
broker router

conf/druid/cluster/query/broker:
jvm.config main.config runtime.properties

conf/druid/cluster/query/router:
jvm.config main.config runtime.properties

``

Are we supposed to re-organize the files in conf directory as per recommeded structure, and then deploy it?

Going forward does Druid recommend to have a single node for coordinator and overlord processes?

Earlier versions had this structure in place already. Further for me, each of the druid cluster is deployed on AWS such that each process runs on a separate EC2 instance (but in 0.15.0 for above I see coordinator and overlord config combined). The reason I am asking is I am upgrading Druid from 0.13.0 to 0.15.0 version.

Thank you in advance.

Regards,

Vinay

Hi,

We just did the same operations on 2 druid clusters (0.12.3 and 0.14.1 to 0.15.0)

And we hit the same point.

The structure present in the archive is to present all the ways to deploy Druid (and I find it cool) but may be disappointing at first sight.

We chose to keep the same structure we had before to not impact our terraform jobs and our habits so much.

Also, considering the bug fixed in 0.15.0 (https://lists.apache.org/x/thread.html/227ed934e5b044d61a7fecd25dcb806763bf1faaa2f5a598ab77d9f1@%3Cdev.druid.apache.org%3E) , you may be interested in postpone your upgrade for a few days (I think)

Guillaume

Thank you Guillaume. Yup its nice, but then at first look it tells me that I will have to make changes to AMI creation script, and move away from current habits of accessing the configuration strcuture. I am also thinking about maintaining the same structure ( downside will be for every new release I will have to make changes to package to match old conf directory structure and then use the custom package for creating am AMI)

Yup, I am aware of the critical bug with 0.15.0 version. Will hold back for few days before going for the upgrade.