You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently we are using configuration file for each role of clusters, in order to customize a cluster, we need to create four files. This approach has a few drawbacks:
It's difficult for user to understand which function toggle to be configured on which file/role
It's difficult to share configuration with other people
It's introducing additional complexity when there is a need to configure one function on each file (remote wal, for example, on both datanode and meta). It's also error-prone.
This situation is also inconsistent with our design philosophy to use one binary for all scenario.
Implementation challenges
Some of the configuration items are highly related to particular instance of the cluster, for example, node_id is only meaning for a datanode instance and cannot share between others. We need a solution for this.
The text was updated successfully, but these errors were encountered:
What type of enhancement is this?
Configuration
What does the enhancement do?
Currently we are using configuration file for each role of clusters, in order to customize a cluster, we need to create four files. This approach has a few drawbacks:
This situation is also inconsistent with our design philosophy to use one binary for all scenario.
Implementation challenges
Some of the configuration items are highly related to particular instance of the cluster, for example,
node_id
is only meaning for a datanode instance and cannot share between others. We need a solution for this.The text was updated successfully, but these errors were encountered: