You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@samza.apache.org by "ASF GitHub Bot (JIRA)" <ji...@apache.org> on 2018/10/16 01:56:00 UTC

[jira] [Commented] (SAMZA-1954) User provided configuration should have higher precedence than generated configuration.

    [ https://issues.apache.org/jira/browse/SAMZA-1954?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16651038#comment-16651038 ] 

ASF GitHub Bot commented on SAMZA-1954:
---------------------------------------

Github user asfgit closed the pull request at:

    https://github.com/apache/samza/pull/728


> User provided configuration should have higher precedence than generated configuration.
> ---------------------------------------------------------------------------------------
>
>                 Key: SAMZA-1954
>                 URL: https://issues.apache.org/jira/browse/SAMZA-1954
>             Project: Samza
>          Issue Type: Bug
>            Reporter: Prateek Maheshwari
>            Assignee: Prateek Maheshwari
>            Priority: Major
>
> With the introduction of Descriptors, there are now two ways to specify configurations - in a file and within the application code.
> This raises the question of what happens if a configuration is specified in both ways with different values. I.e., what is the precedence if a configuration has been set using both the descriptor and in the configuration file.
> Current implementation gives higher precedence to generated configurations than those provided in configuration files.
> This proposal is to make it so that if configuration is present in both, the value in configuration files has higher precedence.
> This has the advantage that operational configurations can be changed at runtime by updating configuration files and redeploying, instead of requiring a full build and deploy. This is the natural/expected behavior for configurations.
> This has the disadvantage that the configuration value in code may not match what's being used at runtime, which may be confusing when debugging. This might be OK since the source of truth for final configuration is the running container's logs anyway (because of various config rewriting and generation stages).
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)