Use Serverless for Apache Spark to Load BigQuery Reviews

Noguchi M. · Reviewed over 1 year ago

Edward K. · Reviewed over 1 year ago

As a beginner in the world of Date Engineering on GCP, I struggle to understand the point of Dataproc in this lab. Also, I wanted to let you new that during the spark execute command I got the following error a couple of times. After the third try, it worked: ERROR: (gcloud.beta.dataproc.batches.submit.pyspark) Batch job is FAILED. Detail: Insufficient 'CPUS' quota. Requested 12.0, available 11.0. Your resource request exceeds your available quota. See https://cloud.google.com/compute/resource-usage. Use https://cloud.google.com/docs/quotas/view-manage#requesting_higher_quota to request additional quota. Running auto diagnostics on the batch. It may take few minutes before diagnostics output is available. Please check diagnostics output by running 'gcloud dataproc batches describe' command.

Andrea T. · Reviewed over 1 year ago

Beini W. · Reviewed over 1 year ago

Ronald Alberto R. · Reviewed over 1 year ago

Antonio B. · Reviewed over 1 year ago

I see opportunities to improve this quick lab: 1) The lab is missing one step which is granting permissions to get the dataproc cluster. I've got the following error message in the VM while executing the batch job: ```` ERROR: (gcloud.beta.dataproc.batches.submit.pyspark) Batch job is FAILED. Detail: Multiple Errors: - Failed to fetch cluster for batch - Permission 'dataproc.clusters.get' denied on resource '//dataproc.googleapis.com/projects/qwiklabs-gcp-00-c467d66f6efc/regions/us-east4/clusters/srvls-batch-1b1a8482-374f-4c44-83d7-6bc417531bed' (or it may not exist). ``` Fortunatelly I was able to figure it out through IAM permissions configuration, but the lab does not provide that guidance. 2) It can be out of scope for a lab, but I think somehow it lacks the explanation for what use cases this solution would be preferred over other ones.

Caio L. · Reviewed over 1 year ago

na

Lipsita N. · Reviewed over 1 year ago

julian c. · Reviewed over 1 year ago

William L. · Reviewed over 1 year ago

"Download these files, click that button, and stuff will happen". That's a summary for this lab. I never worked with Spark before, and I learnt nothing from this lab. It doesn't force you to review what you downloaded, or write you're own transformation (this is an ETL module after all). Additionally, it returned an error the first time I executed the Spark code. With no changes at all, it ran ok the second time.

Pau B. · Reviewed over 1 year ago

faced that max-workers on region lower that submit config

Alexander L. · Reviewed over 1 year ago

Gaurang R. · Reviewed over 1 year ago

Mohamed A. · Reviewed over 1 year ago

Virtualmente Italia s. · Reviewed over 1 year ago

Satyam S. · Reviewed over 1 year ago

Priyanka P. · Reviewed over 1 year ago

nice

NIVASH A. · Reviewed over 1 year ago

Gaetano F. · Reviewed over 1 year ago

nikhil vihar G. · Reviewed over 1 year ago

Yogesh M. · Reviewed over 1 year ago

nikhil vihar G. · Reviewed over 1 year ago

Humberto R. · Reviewed over 1 year ago

CESAR H. · Reviewed over 1 year ago

Norberto G. · Reviewed over 1 year ago