• Azure Storage services Redundancy

    What happens if the storage node on which your blobs are stored fails? What happens if the rack holding the storage node fails? Fortunately, Azure supports something called redundancy. There are four choices for redundancy; you specify which one to use when you create the storage account. You can change the redundancy settings after they are set up, except in the case of zone redundant storage.

     Locally Redundant Storage (LRS) Azure Storage provides high availability by ensuring that three copies of all data are made synchronously before a write is deemed successful. These copies are stored in a single facility in a single region. The replicas reside in separate fault domains and upgrade domains. This means the data is available even if a storage node holding your data fails or is taken offline to be updated.

    When you make a request to update storage, Azure sends the request to all three replicas and waits for successful responses for all of them before responding to you. This means that the copies in the primary region are always in sync.

    LRS is less expensive than GRS, and it also offers higher throughput. If your application stores data that can be easily reconstructed, you may opt for LRS.

     Geo-Redundant Storage (GRS) GRS makes three synchronous copies of the data in the primary region for high availability, and then it asynchronously makes three replicas in a paired region for disaster recovery. Each Azure region has a defined paired region within the same geopolitical boundary for GRS. For example, West US is paired with East US. This has a small impact on scalability targets for the storage account. The GRS copies in the paired region are not accessible to you, and GRS is best viewed as disaster recovery for Microsoft rather than for you. In the event of a major failure in the primary region, Microsoft would make the GRS replicas available, but this has never happened to date.

     Read-Access Geo-Redundant Storage (RA-GRS) This is GRS plus the ability to read the data in the secondary region, which makes it suitable for partial customer disaster recovery. If there is a problem with the primary region, you can change your application to have read-only access to the paired region. The storage client library supports a fallback mechanism Microsoft.WindowsAzure.Storage.RetryPolicies.LocationMode to try to read from the secondary copy if the primary copy can’t be reached. This feature is built in for you. Your customers might not be able to perform updates, but at least the data is still available for viewing, reporting, etc.

    You also can use this if you have an application in which only a few users can write to the data but many people read the data. You can point your application that writes the data to the primary region but have the people only reading the data access the paired region. This is a good way to spread out the performance when accessing a storage account.

     Zone-Redundant Storage (ZRS) This option can only be used for block blobs in a standard storage account. It replicates your data across two to three facilities, either within a single region or across two regions. This provides higher durability than LRS, but ZRS accounts do not have metrics or logging capability.

    Source of Information : Microsoft Azure Essentials Fundamentals of Azure Second Edition


Leave a Reply