It is possible with Log Replay Service. Lean more about LRS here: learn.microsoft.com/en-us/azure/azure-sql/managed-instance/log-replay-service-overview?view=azuresql

@@azuresqlmanagedinstance unfortunately this is not what we were looking for. I ended up having to build our own solution. I am happy to share with Microsoft if I get credit.

​@@Wakko13 do you happen to have the source for this?

​@@petermkaw The script I wrote downloaded the full .bac and differentials .inc files from source system; restore them to a SQL Server container, then export it out as a bacpac file to restore to an Azure Serverless SQL instance (you will need to delete the existing db each day if you do this). Set it up in a daily automated process.

​​@@Wakko13 first off, i appreciate your super quick response. the thing is, our vendor is having a hard time with the restoration, saying you couldn't use log replay to restore database diff backups. basically, a full.bak was done and a diff.bak from the source. then this was given to them for restoration. now i'm no az expert, so my assumption is that they placed this in a blob storage and did all prerequisites for restoration. below is the command to show the current status: vendor says: + PUT FULL & DIFF backups in 1 storage account, restore them at once. => The DIFF backup still being SKIP by AZLRS ------------------------ PS /home/test> az sql midb log-replay show -g rg-testbpp-p-ia --mi sqlmi-testbpp-pia-db01 -n DBtest { "blockReason": null, "currentBackupType": "None", "currentRestorePlanSizeMb": 19, "currentRestoredSizeMb": 19, "currentRestoringFileName": null, "diffBackupSets": [ { "backupSizeMb": 3, "firstStripeName": "DBtest_DifferentialBackup_20240108.bak", "numberOfStripes": 1, "restoreFinishedTimestampUtc": null, "restoreStartedTimestampUtc": null, "status": "Skipped" } ], "fullBackupSets": [ { "backupSizeMb": 19, "firstStripeName": "DBtest_Fullbackup_20240105.bak", "numberOfStripes": 1, "restoreFinishedTimestampUtc": "2024-01-09T17:06:05.716271+00:00", "restoreStartedTimestampUtc": "2024-01-09T17:06:03.791076+00:00", "status": "Restored" } ], "id": "/subscriptions/777777777777/resourceGroups/rg-testbpp-p-ia/providers/Microsoft.Sql/managedInstances/sqlmi-testbpp-pia-db01/databases/DBtest/restoreDetails/Default", "lastRestoredFileName": "DBtest_Fullbackup_20240105.bak", "lastRestoredFileTime": "2024-01-09T17:06:05.716271+00:00", "lastUploadedFileName": "DBtest_DifferentialBackup_20240108.bak", "lastUploadedFileTime": "2024-01-09T08:02:24+00:00", "logBackupSets": [], "name": "Default", "numberOfFilesDetected": 2, "numberOfFilesQueued": 0, "numberOfFilesRestored": 1, "numberOfFilesRestoring": 0, "numberOfFilesSkipped": 1, "numberOfFilesUnrestorable": 0, "percentCompleted": 100, "resourceGroup": "rg-testbpp-p-ia", "status": "Waiting", "type": "Microsoft.Sql/managedInstances/databases/restoreDetails", "typePropertiesType": "LRSRestore", "unrestorableFiles": [] } PS /home/test> i have checked online and seems like the command az sql midb log-replay show -g rg-testbpp-p-ia --mi sqlmi-testbpp-pia-db01 -n DBtest only shows the current status of the log replay job. i asked for the actual command they ran which wasn't provided yet. i just fear that this is a case of wrong (or missing) parameters? or maybe files should be of a certain extension? or is it case sensitive? really appreciate your help dude! *comment edited for sanitation

This is an Azure SQL Managed Instance channel run by SQL MI experts, so take the advice with a grain of salt. Here's documentation on Azure PostgreSQL Flexible Server Start/Stop feature. The doc points you to Azure Portal where you should be able to find the feature. learn.microsoft.com/en-us/azure/postgresql/flexible-server/how-to-stop-start-server-portal

@@azuresqlmanagedinstance Thank you.

Thanks! :)

There is no specific reason why we chose that specific address pool. Example is this video is quite generic, so we were free to chose anything. Depending on other parts of your system and network, you might want to choose address pool that's a subset of addresses you typically use, or you might choose an address pool that does not overlap with some other pool that's already in use.

Managed instances require direct connectivity via VNet-local endpoints for replication traffic. It is, however, possible to establish a failover group listener when both replicas are available via private endpoints, for example in a third virtual network. See scenario 8 in the blog post: techcommunity.microsoft.com/t5/azure-sql-blog/advanced-scenarios-with-private-endpoints-to-azure-sql-managed/ba-p/3902198

You're welcome! :)

Thank you! :)

Thanks for the kind words! We'll add this to the list of ideas for our videos. In the mean time, this table is probably good starting point for understanding performance in general and IOPS specifically: learn.microsoft.com/en-us/azure/azure-sql/managed-instance/resource-limits?view=azuresql#service-tier-characteristics

Short answer - you should rename problematic database and invoke PITR (point in time restore), to re-create database at desired moment T when everything was still fine. With PITR you can choose database name, so you can set it do be "Orders_DB".

Hi John! 1. Yes. 2. One time thing - until you need new storage account/storage container or until SAS token expires (you can set expiration date arbitrarily long). 3. This can be done from SSMS restore wizard as well. Let us know if you have more questions!

@@azuresqlmanagedinstance Thank you!

Thanks Rob for sharing your feedback with us here!

Thank you for the kind words!

Glad you enjoyed it

Estamos de acordo!

Maravilha!

Thank you for the kindest words, Kirby!