Errors checkpointing to S3 for high-scale jobs

[Jamie Grier-2]

I think we need to modify the way we write checkpoints to S3 for high-scale
jobs (those with many total tasks).  The issue is that we are writing all
the checkpoint data under a common key prefix.  This is the worst case
scenario for S3 performance since the key is used as a partition key.

In the worst case checkpoints fail with a 500 status code coming back from
S3 and an internal error type of TooBusyException.

One possible solution would be to add a hook in the Flink filesystem code
that allows me to "rewrite" paths.  For example say I have the checkpoint
directory set to:

s3://bucket/flink/checkpoints

I would hook that and rewrite that path to:

s3://bucket/[HASH]/flink/checkpoints, where HASH is the hash of the
original path

This would distribute the checkpoint write load around the S3 cluster
evenly.

For reference:
https://aws.amazon.com/premiumsupport/knowledge-center/s3-bucket-performance-improve/

Any other people hit this issue?  Any other ideas for solutions?  This is a
pretty serious problem for people trying to checkpoint to S3.

-Jamie

[Stephan Ewen]

For posterity: Here is the Jira Issue that tracks this:
https://issues.apache.org/jira/browse/FLINK-9061

On Thu, Mar 22, 2018 at 11:46 PM, Jamie Grier <[hidden email]> wrote:

> I think we need to modify the way we write checkpoints to S3 for high-scale
> jobs (those with many total tasks).  The issue is that we are writing all
> the checkpoint data under a common key prefix.  This is the worst case
> scenario for S3 performance since the key is used as a partition key.
>
> In the worst case checkpoints fail with a 500 status code coming back from
> S3 and an internal error type of TooBusyException.
>
> One possible solution would be to add a hook in the Flink filesystem code
> that allows me to "rewrite" paths.  For example say I have the checkpoint
> directory set to:
>
> s3://bucket/flink/checkpoints
>
> I would hook that and rewrite that path to:
>
> s3://bucket/[HASH]/flink/checkpoints, where HASH is the hash of the
> original path
>
> This would distribute the checkpoint write load around the S3 cluster
> evenly.
>
> For reference:
> https://aws.amazon.com/premiumsupport/knowledge-
> center/s3-bucket-performance-improve/
>
> Any other people hit this issue?  Any other ideas for solutions?  This is a
> pretty serious problem for people trying to checkpoint to S3.
>
> -Jamie
>