| copyright | lastupdated | subcollection | ||
|---|---|---|---|---|
|
2020-03-03 |
AnalyticsEngine |
{:new_window: target="_blank"} {:shortdesc: .shortdesc} {:codeblock: .codeblock} {:screen: .screen} {:pre: .pre}
{: #config-cluster-cos}
{{site.data.keyword.cos_full_notm}} is a highly scalable cloud storage service, designed for high durability, resiliency and security. See {{site.data.keyword.cos_full_notm}}.
This topic explains how to configure an {{site.data.keyword.iae_full_notm}} cluster to connect to data and applications stored in {{site.data.keyword.cos_short}}.
You should use {{site.data.keyword.cos_full_notm}} as your primary data source and sink as described in Best Practices. Not only the data itself, but also your application or job binaries, for example Spark Python files or Yarn application JARs, should reside in {{site.data.keyword.cos_short}}. By removing all data from the cluster, you make your cluster stateless, which gives you the flexibility to spin up new {{site.data.keyword.iae_full_notm}} clusters as you need them. See Choose the right {{site.data.keyword.cos_short}} Storage configuration.
To work with {{site.data.keyword.cos_short}} in an {{site.data.keyword.iae_full_notm}} cluster:
- Provision an {{site.data.keyword.cos_short}} service instance in {{site.data.keyword.Bluemix_short}}. See Creating a new service instance.
Create an {{site.data.keyword.cos_short}} bucket in the region of your choice and select the configuration, such as resiliency (regional or cross-regional), and storage volume and price. See Creating buckets to store your data.
Make a note of the bucket name that you created. You will need it later when you configure {{site.data.keyword.iae_full_notm}} to work with {{site.data.keyword.cos_short}}.