Rescaling Flink job from an External Checkpoint

[suraj7]

Hi,

I'm using Flink 1.5 with RocksDB State Backend for Checkpointing. According
to the docs, Checkpoints do not support Job rescaling. However, when i tried
to restart a job from an externalised checkpoint with a different
parallelism, the job resumes correctly without any issues.

I would like to know if rescaling a Flink job from an external Checkpoint
based on RocksDB State Backend is reliable or are there any known
issues/drawbacks?

Thanks,
Suraj



--
Sent from: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/

[Chesnay Schepler]

The docs are worded that way since not all backends support it.

I believe rescaling does work for RocksDB checkpoints, but we do not provide any guarantee that this remains to be the case.

Basically, use at your own risk.

On 13.11.2018 13:24, suraj7 wrote:

Hi,

I'm using Flink 1.5 with RocksDB State Backend for Checkpointing. According
to the docs, Checkpoints do not support Job rescaling. However, when i tried
to restart a job from an externalised checkpoint with a different
parallelism, the job resumes correctly without any issues.

I would like to know if rescaling a Flink job from an external Checkpoint
based on RocksDB State Backend is reliable or are there any known
issues/drawbacks?

Thanks,
Suraj



--
Sent from: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/