Develop a convention for the AWS EBS volumes that contain our Solr indexes

[afred]

The EBS volumes that store our Solr indexes are precious to us because they take a long time to build.

So we need some kind of predictable way to manage these particular volumes. Specifically...

As a DevOps I can refer to documented conventions for creating and naming Solr volumes So that they can be easily identified without having to do detective work.

[afred]

First proposal:

Creating a Solr volume

• Naming convention: aapb-solr-{N} where {N} is 1, 2, 3, ...
• Size: Depends on need. Default to as small as possible. You can create new EBS from snapshots of existing EBS, but the new EBS must be at least as large as the snapshot from which it's being created. IOW, you can resize up, but not down.
• Type: General purpose SSD. Could use some more knowledge here, namely:
  • Cost/Performance trade-off of General Purpose SSD over magnetic?
  • Should we do EBS Optimization and use Provisioned IOPS? Cost?
• Availability Zone: Set to same availability zone as the target EC2. Default to us-east-1a.

Attaching a Solr volume to an instance

• Device: Use the default, which will depend on the operating system of the instance to which you are attaching the volume. For instances runing Amazon Linux, it will initially put the device at dev/sdf for the first volume, dev/sdg, dev/sdh, and so on, for additional volumes.