[
https://issues.apache.org/jira/browse/KAFKA-5046?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15962203#comment-15962203
]
ASF GitHub Bot commented on KAFKA-5046:
---------------------------------------
GitHub user kkonstantine opened a pull request:
https://github.com/apache/kafka/pull/2830
KAFKA-5046: Support file rotation in FileStreamSource Connector
- Detect file rotation/removal in FileStreamSource task.
- Extend offset key to include creation time and fileKey attributes.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/kkonstantine/kafka
KAFKA-5046-Support-file-rotation-in-FileStreamSource-Connector
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/kafka/pull/2830.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #2830
----
commit e9f062bc179c9308d683ea590ebcc90854d6c874
Author: Konstantine Karantasis <[email protected]>
Date: 2017-04-09T00:14:47Z
KAFKA-5046: Support file rotation in FileStreamSource Connector
- Detect file rotation/removal in FileStreamSource task.
- Extend offset key to include creation time and fileKey attributes.
----
> Support file rotation in FileStreamSource Connector
> ---------------------------------------------------
>
> Key: KAFKA-5046
> URL: https://issues.apache.org/jira/browse/KAFKA-5046
> Project: Kafka
> Issue Type: Improvement
> Components: KafkaConnect
> Affects Versions: 0.10.2.0
> Reporter: Konstantine Karantasis
> Assignee: Konstantine Karantasis
> Priority: Minor
>
> Currently when a source file is moved (for file rotation purposes, or between
> restarts of Kafka Connect) the FileStreamSource Connector can not detect the
> change, because it only uses the filename as key to its offset tracking.
> Nevertheless, file rotation can be detected easily by checking basic file
> attributes such as the {{fileKey}} in platforms that this attribute is
> supported (for instance file key includes the device id and the inode in unix
> based filesystems) and the file's creation time.
> Such checks need to take place when the task starts and when no more records
> are read during a call to {{poll}}.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
