Spark-on-k8s-operator: Please support release Spark 3.0

Created on 19 Jun 2020  路  6Comments  路  Source: GoogleCloudPlatform/spark-on-k8s-operator

It was released yesterday, can you please update
https://spark.apache.org/docs/latest/

I'm happy to create my own base image if this can already be supported

Most helpful comment

I have just built and pushed the following images:

gcr.io/spark-operator/spark:v3.0.0
gcr.io/spark-operator/spark-py:v3.0.0
gcr.io/spark-operator/spark-operator:v1beta2-1.1.2-3.0.0

All 6 comments

Will create a new release based on Spark 3.0 soon. Stay tuned. Thanks.

I have just built and pushed the following images:

gcr.io/spark-operator/spark:v3.0.0
gcr.io/spark-operator/spark-py:v3.0.0
gcr.io/spark-operator/spark-operator:v1beta2-1.1.2-3.0.0

Thanks @liyinan926, will there be any changes to the API?

@itamar-otonomo There won't be major or backward incompatible changes to the API. There's a couple of new options in Spark 3.0.0 missing in the API. I created https://github.com/GoogleCloudPlatform/spark-on-k8s-operator/pull/954 to add them.

We have pretty much added support for most new options in Spark 3.0.0 that are relevant in the context of the operator. We just added support for a limited form of dynamic allocation without an external shuffle service in #976. Closing this and will prepare a new release for 3.0 soon.

Can you create or point to some docs on the dynamic allocation?

Was this page helpful?
0 / 5 - 0 ratings

Related issues

vivekjain123 picture vivekjain123  路  3Comments

liyinan926 picture liyinan926  路  3Comments

Venkatesh89-Github picture Venkatesh89-Github  路  8Comments

hongshaoyang picture hongshaoyang  路  3Comments

liyinan926 picture liyinan926  路  7Comments