Understanding MapR Object Store Standalone Deployment Mode

Because each MapR Object Store interacts only with one bucket or set of buckets at a time, you can use only Amazon S3 standalone deployment mode.

You cannot have more than one server instance working against a bucket because each MapR Object Store exclusively handles one or more buckets at a time.

To use the S3 API in standalone mode, each MapR Object Store must have its own backend directory in MapR File System. You can map it to either a volume mount point or a directory path.

Each S3 gateway serves an exclusive set of buckets. If a bucket does not exist, then an application can create a bucket through any of the S3 gateways, but it will subsequently only be served through that gateway.

An S3 instance exclusively uses the allocated directory or volume in MapR File System. In case you need to migrate buckets to another S3 instance, move or copy the buckets to another directory or volume.

The following deployment scenario shows one MapR Object Store per cluster that supports multiple applications and multiple buckets with bucket sharing supported.

The following deployment scenario shows two MapR Object Stores for the cluster that supports multiple applications and multiple buckets with bucket sharing supported. In this case, bucket sharing across S3 instances are not supported.

Note: If you want to access multiple buckets without the application having to know about bucket locations beforehand, you can deploy a single MapR Object Store to serve all requests without the need to partition any buckets.