Feature Request - restore SQL database

[fspadaro]

I have a feature request/suggestion regarding the restore SQL Database with override.

My environment

• VM Windows OS - SQL Server application

• Backup Job - Windows Computer (Windows Agent Backup)

• Type: Server Mode: Managed by backup server

• Backup Mode: Entire Computer

• Guest Processing: Enable application-aware processing

• • Transaction logs: Backups logs periodically (backup up logs will be truncated) every 5 minutes

• • Retain log backups: Until the corresponding image-level backup is deleted

• Schedule: every day at 5.00AM

My task
Restore overriding the production database

Using “Veeam Explorer for Microsoft SQL Server” with “Restore point-in-time state to...” option and specifying the restore point with the latest and enable “Perform restore to the specific transaction”, I can select the restore point to the state prior to transaction involving the selected operation.
Executing the restore, an information screen requires to confirm “the override option” and the database is restored successfully.

Next behaviuor
The backup will continue. The transaction log task will continue every 5 minutes without issue.

Veeam will not warn me that no more transaction logs are not consistent with the previous full and so they are unusable and until the next scheduling (ex.: daily in my environment) the database will not be protected by the 5 minutes transaction log backup.

To avoid this it's important, after restore a database with override, to run the primary job to create a new restore point of the database even if to wait the next daily (in may environment) scheduling.

Suggestion
Say that, I think it's important to indicate with "screen message" to execute a new restore point of the backup to restart to manage the new transaction logs.

Federico

[Andreas Neufert]

Basically the feature request is that the restore wizard will give you a visible note that you need to run a new (image) level backup in order to protect the restored database?

[Andreas Neufert]

@PetrM

[fspadaro]

Basically the feature request is that the restore wizard will give you a visible note that you need to run a new (image) level backup in order to protect the restored database?

Correct, to protect the restored database and all subsequent transaction logs

[PetrM]

Hello,

By design, we skip the restored database from T-log processing and do not touch logs of such database before the next successful image-level run.

@fspadaro
Have you checked the transaction log backup statistics? There should be a message indicating that the database was skipped.

Thanks!

[fspadaro]

Hello @PetrM,

for skipping the restored database from T-log processing by design is there a document? Because I don't find anything

In the statistics windows I find this line:
26/09/2024 09:53:13 :: Backed up 521.5 KB of transaction logs for 3 databases: 20240911_1704;model;PROVA at 27.4 KB/s with bottleneck: Target (Network)
26/09/2024 09:58:15 :: Backed up 435.0 KB of transaction logs for 2 databases: model;20240911_1704 at 24.2 KB/s with bottleneck: Target (Network)
.......................................
26/09/2024 10:08:25 :: Performing SQL Server transaction log backup for model;20240911_1704
as you said, the database restored has been skipped, but there isn't a clear message that the database will be skipped.

A customer did not look at the statistics window and did not notice that the database had been skipped.

So, I think it important to warn with screen message this behaviour, don't you believe?

Thanks

[Andreas Neufert]

I am not sure if this is really an issue and it would be an issue throughout the whole product.

Whenever you restore a server/database/files/... you need to perform a new backup.
Specifically, the DBAs understand that the log backups are useless without an initial backup.

So the feature request should maybe be a popup for all restores (not only SQL restores) that say, that you need to perform a new backup / backup run in order to protect the restored data?

[fspadaro]

Hello @Andreas Neufert,

exactly, as you said:

So the feature request should maybe be a popup for all restores (not only SQL restores) that say, that you need to perform a new backup / backup run in order to protect the restored data?

In my request I say to add or integrate on the override popup an indication to perform a new backup.

[Andreas Neufert]

@PetrM Please add to the product request tracker and check on wizards.

Federico thanks for the input. We will consider this.

[PetrM]

Hello,

The request is noted.

In the meantime, I would like to clarify one detail:

"The restored database has been skipped, but there isn't a clear message indicating that the database will be skipped."

We have checked this case in our lab and we have observed the following message in the T-log backup job statistics window: "Skipping deleted database: DBName". We don't know about the restored database before the next image-level run.

@fspadaro could please check the log backup job statistics and open a support ticket if you don't see it? If you proceed with the support case, please share a support case ID.

Thanks!

[fspadaro]

Hello PetrM,

The customer case I closed after understanding this behavior, indicating the procedure to be performed in case of full restore.

To understand it better and do various tests I replicated the problem in my lab but I don’t think I can open a new case for my test environment.

Thanks!

[PetrM]

Hi Federico,

"Please try to open a case anyway. If it is closed due to any environment/license limitations, just send me a private message and we'll discuss the problem. Maybe there are some corner cases that need to be researched, but in general, the behavior by design is to show the corresponding message in T-log backup statistics.

Thanks!