Hi!

Checkpoints in Flink need to go to a file system that is accessible across
machines. Otherwise there could be no recovery of a data of a failed
machine.

The cleanup is also triggered by a different node than the node that
checkpoints - hence you see no cleanup in your setup.

Best,
Stephan

On Fri, Sep 8, 2017 at 6:19 PM, rnosworthy <[email protected]> wrote:

> Thanks for the response.
>
> Thats correct, they do not get purged/deleted while the job is running. I
> have 3 concurrent jobs running and there are 3 directories in the data
> directory.
>
> /var/data/flink/2375c69006bfeca9644171f31b444dff
> /var/data/flink/c3264bb6d5e068d6440bbb21069b7d28
> /var/data/flink/f81d50eb4644cdf65f8f0513713c9d61
>
> in each one of those folders there is chk-1 all the way to chk-2777 for
> example
>
> I do not have a hdfs, I have the flink task manager on a debian vm
>
>
>
> --
> Sent from: http://apache-flink-user-mailing-list-archive.2336050.
> n4.nabble.com/
>

Reply via email to