Spark worker memory doesn't set executor memory to maximum

[icarosadero]

Name and Version

bitnami/spark:3.5.2

What architecture are you using?

amd64

What steps will reproduce the bug?

Simply start with the defaults and set SPARK_WORKER_MEMORY=1G to some other valuer other than 1G.

What is the expected behavior?

It is expected that the executor memory will also be updated to the value set in SPARK_WORKER_MEMORY

What do you see instead?

In the console, it can be seen that the executor memory is still the dafault of 1G

Additional information

Passing SPARK_EXECUTOR_MEMORY in the docker compose doesn't do anything.

[carrodher]

Thank you for bringing this issue to our attention. In fact, it seems there is something wrong since the env. variable defined is SPARK_WORKER_MEMORY but the one used in the entrypoint is SPARK_EXECUTOR_MEMORY, see

$ ag 'SPARK_.*_MEMORY' bitnami/spark/3.5/debian-12/
rootfs/opt/bitnami/scripts/spark/entrypoint.sh
66:      "-Xms${SPARK_EXECUTOR_MEMORY}"
67:      "-Xmx${SPARK_EXECUTOR_MEMORY}"

docker-compose.yml
21:      - SPARK_WORKER_MEMORY=1G

Since you discovered the issue, if you're interested in contributing a solution, we welcome you to create a pull request. The Bitnami team is excited to review your submission and offer feedback. You can find the contributing guidelines here.

Your contribution will greatly benefit the community. Feel free to reach out if you have any questions or need assistance.

[icarosadero]

So, after some extra search on the internet, I discovered that you can request extra executor memory from the client. By changing the spark.executor.memory parameter. In Scala, that would look like so:

val spark = SparkSession.builder.master("spark://spark:7077").config("spark.executor.memory", "10g").getOrCreate()

It isn't a definitive solution, but might help with most use cases.

[github-actions]

This Issue has been automatically marked as "stale" because it has not had recent activity (for 15 days). It will be closed if no further activity occurs. Thanks for the feedback.