Build a kubernetes environment for running examples on Spark runner
- Dominant language
- Java
- Stars
- 8.7k
- Forks
- 4.7k
- Avg merge
- 2d 2h
- Merged PRs (30d)
- 205
Description
This will help getting examples setup and deployed quickly in a multi-node Spark standalone cluster.
This tutorial can be used as a guideline:
https://tensorflow.github.io/serving/serving_inception
Some work going on in k8s sig-big-data that could be built-on:
https://docs.google.com/document/d/1pnF38NF6N5eM8DlK088XUW85Vms4V2uTsGZvSp8MNIA/edit#heading=h.i8o3am52bhuk
Imported from Jira [BEAM-1536](https://issues.apache.org/jira/browse/BEAM-1536). Original Jira may contain additional context.
Reported by: nlamba.
Contributor guide
Research direction
Start with the linked TensorFlow Serving tutorial and the k8s sig-big-data document, then review the Apache Beam Spark runner examples and deployment entry points. Done means providing a Kubernetes environment that lets examples be set up and deployed quickly on a multi-node Spark standalone cluster.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- kubernetes, spark
- Domain
- distributed-systems, infrastructure
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100