连接器保留其从文件读取的最后位置"的选项卡，因此在读取文件时崩溃的情况下，连接器可以从中断处继续.

问题是您删除文件时未将偏移量重置为0，因此它基本上看不到任何新数据，因为它等待新数据从头开始显示特定的字符数... >

解决方法，如果要重置偏移量.如果您以独立模式使用connect，则默认情况下，偏移量存储在/tmp/connect.offsets中，只需从那里删除它们即可.

从长远来看，我们需要一个更好的文件连接器:)

I was following step #7 (Use Kafka Connect to import/export data) at this link:

http://kafka.apache.org/documentation.html#quickstart

It was working well until I deleted the 'test.txt' file. Mainly because that's how log4j files would work. After certain time, the file will get rotated - I mean - it will be renamed & a new file with the same name will start getting written to.

But after, I deleted 'test.txt', the connector stopped working. I restarted connector, broker, zookeeper etc, but the new lines from 'test.txt' are not going to the 'connect-test' topic & therefore are not going to the 'test.sink.txt' file.

How can I fix this?

The connector keeps tabs of its "last location read from a file", so in case it crashes while reading the file, it can continue where it left off.

The problem is that you deleted the file without resetting the offsets to 0, so it basically doesn't see any new data since it waits for new data to show starting at a specific character count from the beginning...

The work-around if to reset the offsets. If you are using connect in stand-alone mode, the offsets are stored in /tmp/connect.offsets by default, just delete them from there.

In the long term, we need a better file connector :)