connect-dsitributed process dies if null value goes into incrementing.column.name is null about kafka-connect-jdbc HOT 5 CLOSED

Just to make sure I understand, you'd prefer it if the connector would stay alive (but at error state) and keep checking for new configuration?

from kafka-connect-jdbc.

Yup,
Otherwise, we would need to restart "connect-distributed" each time an error occurs even while setting up / adding a connector. If there is stronger error checking with the REST API, recovering and taking corrective actions would be possible. The standard HTTP codes for REST errors could be sent back to let the app developer know that the call has failed. I hope that makes sense

from kafka-connect-jdbc.

@joshivinay Thanks for reporting this issue. We made some improvements of Kafka Connect in CP 3.0 release that supports status tracking of connector and tasks. In case that a connector is not correctly initialized and started, the connector will transit to FAILED state, however, the exceptions during connector start are properly caught. As a result, the Worker is not crashed you should be able to reconfigure the connector with correct configurations.

CP 3.0 also provides REST API report status of connectors and tasks. You should be able to use that API to inform the App of the connector status.

from kafka-connect-jdbc.

Okay,
We will check it out with the CP 3.0 release.

Thanks
Vinay Joshi
Kwartile, Inc.
Email: [email protected] [email protected]

On Thu, Jun 2, 2016 at 12:15 PM, Liquan Pei [email protected]
wrote:

@joshivinay https://github.com/joshivinay Thanks for reporting this
issue. We made some improvements of Kafka Connect in CP 3.0 release that
supports status tracking of connector and tasks. In case that a connector
is not correctly initialized and started, the connector will transit to
FAILED state, however, the exceptions during connector start are properly
caught. As a result, the Worker is not crashed you should be able to
reconfigure the connector with correct configurations.

CP 3.0 also provides REST API report status of connectors and tasks. You
should be able to use that API to inform the App of the connector status.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#57 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AKzV2J3_K-_ecfjJbYYhlxtMb-LQw6Iaks5qHnv0gaJpZM4H5Q1M
.

from kafka-connect-jdbc.

Connect in distributed mode should not crash if a connector fails. There were improvements in 3.0 that @Ishiihara mentioned. In 3.1, there are further robustness improvements with https://issues.apache.org/jira/browse/KAFKA-4042.

from kafka-connect-jdbc.