Azure storage accounts provide a namespace in which to store data objects. These objects could be blobs, file, tables, queues and virtual machine disks. This post focuses on the pieces necessary to create a new storage account for use within Azure Resource Manager portal.
To setup a storage account go to the Azure Resource Manager Portal, select storage accounts and then click the “Add” button. From there you’ll have some familiar settings that will need to be filled out such as a unique name for the account, a subscription to use for billing, a resource group for management, and a location for the region to be used. The rest of this article explains the additional settings shown in the screenshot below.
In the “Account Kind” section, there are two types of storage available, General Purpose and Blob. I found this option to be incredibly confusing because the general purpose storage also supports blob storage as well. Blob storage accounts are for applications requiring block or append type blob storage and don’t support page blobs. The blob storage referred to in the “General Purpose” type are page blobs and provide persistent block storage similar to the Amazon’s EBS.
If your goal is to setup storage for a virtual machine use, then you’re looking for General Purpose storage for the account kind property. I actually found that I was able to select a blob storage account when deploying a virtual machine, but the deployment failed and the error message shows that the storage account type was not supported.
When you drill down into your storage account you’ll notice that there are services listed for the storage account and you can drill into each service to get more information.
If we looked at the blob service, we’ll see a container named “vhds” which is where the Azure machine virtual disks are stored. You’ll need to go into this container to delete the vhd when you remove the virtual machine from Azure unless you are removing the entire resource group.
Performance – General Purpose Only
When you create a storage account you must pick a performance tier and this can’t be modified afterwards so choose wisely. There are two types of performance as of the time of this writing.
Standard – Backed by magnetic disks and provide lower cost per GB. These are best used for applications requiring bulk storage with infrequently accessed data.
Premium – Backed by solid state drives offering low latency performance. Recommenced for I/O intensive applications such as databases.
In the replication box there are several options. These options are dependent upon the “Account Kind” and “Performance” selections that you’ve selected earlier. There are four main types of replication that can be used in your Azure storage account.
Locally Redundant Storage (LRS) – This type of storage replicates data within the region where your storage account is created. Every data request in the storage account is replicated three times in different fault domains.
Zone Redundant Storage (ZRS) – This storage replicates data across two or three facilities within a region or across regions. ZRS provides protection against a facility failure.
Geo-redundant Storage (GRS) – Replicates data to a second region that is hundreds of miles away and provides protection from regional outages. For example GRS would have the East US region replicate to the West US region to provide regional durability. It would not have East US replicate to East US 2 because they are geographically located too close together. GRS maintains six copies of the data, three within each datacenter. The data in the secondary location is replicated asynchronously meaning that it is eventually consistent. The recovery point objective on GRS is about fifteen minutes but there is no published SLA on this.
Read-access Geo-redundant Storage (RA-GRS) – This storage is similar to the Geo-Redundant Storage described earlier except that it allows for read-only access at the secondary location.
Access Tier – Blob Storage Only
If you’ve selected “Blob storage” as part of your Account kind, you’ll need to select an access tier. There are two different tiers that can be chosen from. A hot tier for data that is frequently accessed and a cool tier for data that is infrequently accessed. The cool tier provides a cost savings but also provides less performance.