Re: Behavior of the cancel command

Thanks for looking into this, Jürgen! I opened a Jira issue: https://issues.apache.org/jira/browse/FLINK-6427 > On 29. Apr 2017, at 09:15, Jürgen Thomann > wrote: > > Hi Aljoscha, > > In my case the valid-length file created contains a value

Re: Behavior of the cancel command

Hi Aljoscha, In my case the valid-length file created contains a value which e.g. says 100 MB are valid to read for exactly once but the file with the data is only 95 MB large. As I understand it the valid-length file contains the length of the file at the checkpoint. This does also not happe

Re: Behavior of the cancel command

Hi Jürgen, Is there missing data with respect to what should have been written at the time of the cancel or when the last checkpoint (or in that case, the savepoint) was performed. I’m asking because the cancel command is only sent out once the savepoint has been completed, as can be seen at [1]

Behavior of the cancel command

Hi, I had some time ago problems with writing data to Hadoop with the BucketingSink and losing data in case of cancel with savepoint because flush/sync command was interrupted. I tried changing Hadoop settings as suggested but had no luck at the end and looked into the Flink code. If I unders