kafka-jira mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sönke Liebau (JIRA) <j...@apache.org>
Subject [jira] [Commented] (KAFKA-5563) Clarify handling of connector name in config
Date Fri, 17 Nov 2017 09:47:00 GMT

    [ https://issues.apache.org/jira/browse/KAFKA-5563?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16256744#comment-16256744
] 

Sönke Liebau commented on KAFKA-5563:
-------------------------------------

As KAFKA-4930 is approaching a stable state I looked at this again and found that a check
like this is already implemented for [updating a connectorconfig|https://github.com/apache/kafka/blob/e31c0c9bdbad432bc21b583bd3c084f05323f642/connect/runtime/src/main/java/org/apache/kafka/connect/runtime/rest/resources/ConnectorsResource.java#L135]


I've prepared a pull request that moves this check into a small helper function and call this
from the createConnector method as well. 

I've also tested merging this with the PR for 4930 which works nicely (for now), if any problems
occur I'll rebase whichever PR gets merged later.

> Clarify handling of connector name in config 
> ---------------------------------------------
>
>                 Key: KAFKA-5563
>                 URL: https://issues.apache.org/jira/browse/KAFKA-5563
>             Project: Kafka
>          Issue Type: Bug
>          Components: KafkaConnect
>    Affects Versions: 0.11.0.0
>            Reporter: Sönke Liebau
>            Assignee: Sönke Liebau
>            Priority: Minor
>
> The connector name is currently being stored in two places, once at the root level of
the connector and once in the config:
> {code:java}
> {
> 	"name": "test",
> 	"config": {
> 		"connector.class": "org.apache.kafka.connect.tools.MockSinkConnector",
> 		"tasks.max": "3",
> 		"topics": "test-topic",
> 		"name": "test"
> 	},
> 	"tasks": [
> 		{
> 			"connector": "test",
> 			"task": 0
> 		}
> 	]
> }
> {code}
> If no name is provided in the "config" element, then the name from the root level is
[copied there when the connector is being created|https://github.com/apache/kafka/blob/trunk/connect/runtime/src/main/java/org/apache/kafka/connect/runtime/rest/resources/ConnectorsResource.java#L95].
If however a name is provided in the config then it is not touched, which means it is possible
to create a connector with a different name at the root level and in the config like this:
> {code:java}
> {
> 	"name": "test1",
> 	"config": {
> 		"connector.class": "org.apache.kafka.connect.tools.MockSinkConnector",
> 		"tasks.max": "3",
> 		"topics": "test-topic",
> 		"name": "differentname"
> 	},
> 	"tasks": [
> 		{
> 			"connector": "test1",
> 			"task": 0
> 		}
> 	]
> }
> {code}
> I am not aware of any issues that this currently causes, but it is at least confusing
and probably not intended behavior and definitely bears potential for bugs, if different functions
take the name from different places.
> Would it make sense to add a check to reject requests that provide different names in
the request and the config section?



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message