Search the docs

Planning Your Hammerspace S3 Deployment

Hammerspace supports running multiple S3 servers, each with a unique configuration. The servers are hosted on all the nodes running the S3 Service, which allows each S3 Server to be a scale-out entity.

The following tasks describe a generic workflow for configuring S3 storage for a Hammerspace infrastructure.

  1. Identify AD S3 users or create local Hammerspace S3 data accounts

  2. Create an S3 server

  3. Setup the S3 server DNS records

  4. Setup S3 bucket containers and buckets

  5. Manage users and access

  6. Configure S3 clients

Prepare for S3 Configuration

Several data points should be decided upon before configuring the S3 server.

  • Endpoint Name: Each S3 server is usually addressed using an endpoint name, which determines which S3 server the traffic is routed to. This name is represented by a DNS record, which is recommended to be configured before setting up the S3 server.

  • Data Access: S3 buckets have two methods for data access, traditional path-based or virtual-host style. It is essential to know that the virtual-host style requires additional DNS entries for each bucket accessed using virtual-host addressing.

  • Bucket Management: Hammerspace supports managing buckets with the administrative API, creating and deleting buckets via the S3 API, creating or deleting buckets within bucket containers by SMB, NFS, or S3 clients, or a combination of all three.

  • Existing Data Access: If sharing existing data (an entire share or a sub-folder within a share), determine what kind of access should be provided via S3, as permissions are standard across all protocols. However, access can be limited via S3 using an ACL.

Prerequisites

Hammerspace S3 servers support connections using the same DSX floating IP addresses (FIPs) used by SMB and NFS clients. To help balance S3 traffic and ensure that clients can maintain S3 connections during maintenance or unexpected outages of DSX servers, you should create one DNS A record for each DSX FIP, for each S3 server that you create, just as you did for the cluster itself.

The following image shows a typical DNS configuration for a Hammerspace cluster with two DSX nodes and configured with two S3 servers.

s3 prerequisites image1
Figure 1. Prerequisites
While the S3 servers also support connections over the DSX static IP addresses, those IPs should not be used since they cannot migrate to other DSX nodes in the event of DSX node maintenance or an outage.

Supported S3 API Operations

Hammerspace supports a subset of the S3 API covering bucket, object and multipart-upload operations, along with path-style and virtual-hosted bucket addressing and presigned URLs using AWS Signature Version 4 query parameters.

The operation list is release-specific, so it is maintained in one place rather than duplicated here. For the authoritative list for your release — including the operations that are accepted but behave in a fixed way, such as the Intelligent-Tiering calls — see the S3 API Overview section of the Hammerspace Administration Guide.

Operations that do not appear in that list are not supported.