Ask HN: What are the alternatives of hosting Apache Spark? | Hacker News Reader

Ask HN: What are the alternatives of hosting Apache Spark?

28 pointsmuramira17 comments

I truly love what databricks is doing, but their pricing model is unpredictable. Are there any other hosting companies that provide a fixed price?

View on HN

If your spark jobs are mostly batch workloads, that can tolerate moderately infrequent failures and restarts, try using google dataproc with preemptible vms or amazon emr using spot instances.

Depending on your use case, you might spend many times less than you would using regular VMs. Many instances that are several dollars an hour on AWS can be used for a fraction of the price.

Its also fairly easy to automate the region selection and bid (on AWS that is, not sure about gcloud).

If you need streaming, obviously this might not be the way to go.

I find the EMR markup to be substantial; if I weren't working in a corporation, I would stand up my own spark clusters, e.g. using spark-ec2

It is, but you can run the bare minimum number of core nodes (3 I think?) and use spot instances for any others.

At a previous job, we just built our own ec2 image that ran spark in standalone mode for ephemeral spark clusters, and it was wonderful and cheap. And the clusters launched very fast compared to EMR.

Rewrite all of your jobs using Apache Beam. Then use whatever runner you want: Spark, Flink, Google Cloud Dataflow, etc.

Google Dataproc - very good and very soon they will release kubernetes as the manager instead of yarn.

"very soon they will release kubernetes as the manager instead of yarn"

Do you have more details on this?

https://cloudplatform.googleblog.com/2018/03/learn-to-run-Ap...

Check out AWS Glue: https://aws.amazon.com/glue/

Disclosure: I work on this service

So I've been knee deep in glue for the past month. Great service but we really need more examples and an easier dev flow. I wasn't able to get consistent results between juypter and dev endpoint as i was with the glue running the script directly. So i spent a lot of wasted time waiting for glue to run my job that eventually failed and then getting a less than helpful error message. Now that i have my jobs running and orchestrated with step functions it's a thing of beauty but should have taken 1/10 of the time if i had good examples and a proper environment.

All 3 major cloud providers have offerings in this space. Amazon [0], Google [1], Microsoft [2].

[0]: https://aws.amazon.com/emr/

[1]: https://cloud.google.com/dataproc/

[2]: https://azure.microsoft.com/en-us/services/databricks/

You could give AWS EMR a shot, it probably doesn't offer as much as databricks but should have consistent pricing.

Run Spark on a managed Kubernetes like GKE? There is experimental support for using Kubernetes as the cluster manager.

https://apache-spark-on-k8s.github.io/userdocs/index.html

You can try Qubole [0]. The pricing is a small percentage of what you pay to the cloud provider, so it's predictable to an extent.

[0]: https://www.qubole.com/

Disclosure: I work here.

https://hortonworks.com/products/data-platforms/cloud/aws/

We're using Spark on EMR with Data Pipeline to do ETL and to run Scheduled Jobs. Data pipelines terminates the cluster once ETL or job gets completed, helps us a lot to save cost.

what part of it is unpredictable? I guess if you know how much VMs or EC2 you're planning on using its the same pricing model as Dataproc or EMR

Check out Azure Databricks

Reply on news.ycombinator.com