1
votes

We have created a Storage API which will be called by a front end application to retrieve data from Azure Blob Storage Account. The Storage Account is in Canada Central using GZRS replication and is using Private Endpoint. So now we have to do a DR testing to ensure that the Storage Account fails-over from the Primary Region to the Secondary Region and the Storage API connects to the Secondary Region and gets the data from the Secondary Region instead of the Primary. So if we do a user initiated failover in the portal by following the article https://docs.microsoft.com/en-us/azure/storage/common/storage-initiate-account-failover?toc=%2Fazure%2Fstorage%2Fblobs%2Ftoc.json&tabs=azure-powershell

will the Storage API automatically route the traffic to the Secondary Region once Primary is down? We have a connection string only defined in the code which has the Private Endpoint IP

Also, i have another question related to this.


Are Secondary Regions for Storage Accounts in case of GRS or GZRS are taken the paired region and is that something , the user cannot choose on their own, is that correct?

1

1 Answers

0
votes

To answer your first question : From the microsoft docs : When failover occurs, the secondary region becomes your new primary region, and all data is then accessible from this new primary region. All DNS records, which relate to your storage account, have their DNS endpoints updated to point to the new primary region. This redirection requires no changes to your application code. Link to above doc : https://docs.microsoft.com/en-us/learn/modules/ha-application-storage-with-grs/4-design-applications-ha-dr

But you should implement the circuit breaker pattern in your application so that the application once again reads from the Primary region when it is up and running again. For an example of implementing the circuit breaker pattern in your application see : https://docs.microsoft.com/en-us/learn/modules/ha-application-storage-with-grs/5-exercise-deploy-ha-application

And to answer your second question : The user cannot choose the secondary region. Azure automatically uses the paired region for GRS/GZRS. See : https://docs.microsoft.com/en-us/learn/modules/ha-application-storage-with-grs/2-storage-account-replication-features

Geo-zone-redundant storage (GZRS) combines the high availability benefits of ZRS with GRS. With this replication type, your data is copied across three availability zones in one region. Data is also replicated three times to another secondary region that's paired with it. This way, your zone-redundant data is also secure from regional level outage.