Azure Managed Instance and Failover / DR Options - azure

Azure SQL Managed Instance became generally available on 1-October, 2018. However, I've seen virtually nothing written about what options we have with regards to fail-over and disaster recovery. Specifically I would like to know what happens in the event of a server failure hosting an MI instance. I would also like to know what happens in the event of a data center failure. Does anyone have information about this?

Currently, Active Geo-replication is not available in Managed Instance.
Note: Active Geo-replication support is coming to public preview very soon.
For more details, refer "Overview:Active geo-replication and auto-failover groups" and "Setup DR with SQL Managed Instance"

Update: Geo-replication for Azure SQL Managed Instance is now in public preview.
For more details, see https://learn.microsoft.com/en-us/azure/sql-database/sql-database-auto-failover-group#best-practices-of-using-failover-groups-with-managed-instances

Related

Azure free account - does it cover SQL Managed Instance?

I have started using my free Azure account and I found out that I cannot create SQL Managed Instance. I get a cryptic error message telling me to change subscription or region, no clear information. The list of free services does not include SQL MI but it does not mean much. SQL Dedicated Pool or Synapse are also not listed but I tried to create them and the Portal does not complain yet even though I did not click the final Create button yet.
So SQL Managed Instances are only available on certain subscription types. See:
You probably have an Azure Trial subscription. If not, you might also want to check your region as there are region limitations as mentioned in the article above.

Convert Premium To Standard

We have a VM using premium managed drives that is also replicated to another azure data center using azure site recovery. I am aware of how to convert the premium drives to standard by deallocating the vm and changing the drive type. However I suspect I will need to stop and remove the disaster recovery replication and reinitializing vm replication resulting in the loss of all previous recovery points.
Does anyone know for sure and what the process to convert the disks given VM replication would be.
Thx.
You need to reach out to the right Azure Support Team. The Azure Site Recovery support team should be able to give you the correct information, they handle disaster recovery replication scenario.
You want to make sure you are getting vetted information on issues like this.

Does enrolling a SQL Azure database in Geo-Replication provide failover out of the box?

If I go into the Azure portal and go to a SQL Azure db and click on Geo-Replication I can select another data center to have a secondary database in. I can configure this as "readable." With that done, do I automatically get failover?
So for example, if my primary db is in Central US and I configure Geo-Replication to US East 2, will anything automatically failover my db to US East 2 if there is an error in Central US? Or do i have to initiate the failover through the portal or some code/monitoring solution? And would i have to update my connection string or does the azure infrastructure manage this for me?
I've reviewed a few docs below about this but looking for some more input:
- https://azure.microsoft.com/en-us/documentation/articles/sql-database-designing-cloud-solutions-for-disaster-recovery/?rnd=1
- https://azure.microsoft.com/en-us/documentation/articles/sql-database-geo-replication-failover-portal/
- https://azure.microsoft.com/en-us/documentation/articles/sql-database-geo-replication-overview/
https://azure.microsoft.com/en-us/documentation/articles/sql-database-geo-replication-overview/
do i have to initiate the failover through the portal or some
code/monitoring solution?
Yes, you have to initiate the fail over explicitly. There is no automatic failover in case the primary goes offline.
would i have to update my connection string or does the azure
infrastructure manage this for me?
You would have to update connection string explicitly as well.
FailOver and DR drill sections of this link should provide necessary info, it also talks about keeping firewall rules and users in sync between primary and secondary : https://azure.microsoft.com/en-us/blog/spotlight-on-new-capabilities-of-azure-sql-database-geo-replication/

It is possible use the same sql azure instance from two different cloud service of two different subscription?

I have one Microsoft Azure subscription with one cloud service and one sql azure instance. Now I want create another cloud service with a different subscription (using a different microsoft account). With this second cloud service, can I use the same sql azure instance of the first subscription? (I need to share data between the two cloud service)
Or there may be performance issues?
Thanks in advance
Yes. Azure SQL DB instance can be accessed from different subscription as long as you have the connection string, username and password to the Azure SQL instance. As long as both the services are from the same region, there is no performance issue.
Yes, sure. From user perspective SQL Azure is mostly an ordinary SQL Server which you can access from anywhere in the world (given that the firewall rules allow that access) - from Azure services, from VMs in some other services hosted elsewhere, from your desktop, from servers in your company server room.
Network latency might kick in. Also more clients to the same instance mean more load. Also there's a limit on number of concurrent connections. Other than that - no problems.
You need to make sure are a member in each Azure instance to be able to use the others SQL DB

Long Term Support for Azure

What guarantee does Microsoft give to providing long term support for Azure? If Microsoft was to shutdown Azure how long would they keep the Azure cloud up and running? Has anyone regretted using a SQL Azure feature as it harmed their ability to move off of SQL Azure?
OK. A few questions there. Some can really only be answered officially by Microsoft but I'll take a stab at providing at least some detail for you.
1. What guarantee does Microsoft give to providing long term support for Azure?
Microsoft commit to providing at least 12 months notice for any disruptive change. This is set out in their Online Services Support Lifecycle document. http://support.microsoft.com/gp/OSSLpolicy
2. If Microsoft was to shutdown Azure how long would they keep the Azure cloud up and running?
Per the above. I would consider that a disruptive change and expect them to provide a minimum of 12 months notice.
3. Has anyone regretted using a SQL Azure feature as it harmed their ability to move off of SQL Azure?
There are very few features that are only available in SQL Azure. IN terms of shipping features I can only think of Federations off the top of my head. It's a unique feature in that it's only somewhat interesting for on-premise deployments as you don't typically have elastic capacity on tap on premise and you can probably take other approaches such as a monolithic DB server + storage partitioning to solve your problems. In short I haven't had such regrets.

Resources