How long to wait on upgrade to V12?

[garth1138]

My upgrade from V11a with latest patch to V12 latest (On a Windows 2012 R2 Server) has been on Step 1 for an hour. How long should I wait?
Can I kill the upgrade process and start over?

Last write to C:\ProgramData\Veeam\Setup\Temp\BackupSrvLog.log

***Veeam*** SRV: [13.03.2023 14:09:15]:<00000004> [VEEAM] INFO [MSSQL] Column [Backup.Model.Backups.ArchiveInfo].[archive_index_id] has been successfully renamed to [legacy_per_vm_archive_index_id]
***Veeam*** SRV: [13.03.2023 14:09:15]:<00000004> [VEEAM] INFO [MSSQL] Column [dbo].[Backup.Model.Backups.ArchiveInfo].id has been successfully dropped
***Veeam*** SRV: [13.03.2023 14:09:15]:<00000004> [VEEAM] INFO [MSSQL] Database version has been incremented to <8727> (ContentType=[Schema])
***Veeam*** SRV: [13.03.2023 14:09:15]:<00000004> [VEEAM] INFO [MSSQL] Database version has been incremented to <8728> (ContentType=[Schema])

This was almost an hour ago and nothing is moving.

Any recommendations?

[garth1138]

Found SQL Server stuck with error "Timeout occurred while waiting for latch: class 'METADATA_SEQUENCE_GENERATOR'"
Tried to Stop SQL Server with SQl Configuration Management console. No luck. It is hung. So I force killed the SQL Server Instance by killing the PID.
Veeam then complained that it could not finish install due to lost connection with SQL Server. "Fatal error during installation."

C:\ProgramData\Veeam\Setup\Temp\BackupSrvLog.log seems to show install rollback. However, since SQL was shutdown abnormally, not sure what will happen. Will reboot server and try again....

[garth1138]

After reboot, Veeam12 Setup said I needed to remove previous installation. Uggh... Which meant removing all Veeam and starting from scratch.
I will need to import my config backup after install. However, new v12 install is installing PostgreSQL. I don't mind, but hopefully it will "just work"

[Mildur]

Hi Garth

You won‘t be able to import a v11 configuration backup to a V12 postgresql server. Only configuration backups created by v12 will work.

If a new try is not successfull, please provide a case number for this technical issue. We cannot fix technical issues over the forum.

Best,
Fabian

[garth1138]

Stickelbats.
Well, you are correct on that one Fabian. Pointed out right in the restore window when trying.
So, I am removing the V12 and will re-install V11, apply the config backup, and try a new V12 upgrade.
I'll post results....

[garth1138]

Results:
Removed all Veeam and reinstalled 11a with latest patch. New install created a new SQL instance (2016)
Restored config backup to get server back to where it was.
All backup jobs ran fine.
Started upgrade again.
All went fine this time. Upgrade took approx 20 minutes.

Happy camper

Will look at migrating to PostgreSQL in a couple days.

[DanielJ]

It might be interesting to know that I have the exact same problem - last message is "Database version has been incremented to <8728>", then nothing. And SQL Server has logged the same message about "Timeout occurred while waiting for latch: class 'METADATA_SEQUENCE_GENERATOR'". I had to do other stuff at the same time, so over three hours have passed without progress. Maybe others have the same issue as well? Could be worth investigating. I do not look forward to having to do a full reinstall just to be able to attempt the same upgrade again.

[Mildur]

@garth1138
Thanks for the update

@DanielJ
Instead of doing a full reinstall, you can always open a support case (severity 1 if you must full fill SLA's).
We need your logs to investigate this behavior.
Please provide me with the case number so I can follow the case from my side.

Best,
Fabian

[DanielJ]

Ticket with setup logs: 05944938

[bjdboyer]

I've always found rebooting the Veeam server(s) including SQL if a separate server BEFORE doing any Veeam upgrades helps. My one and only time an upgrade failed was due to a Windows issue that was solved just by rebooting. All my Veeam servers except for repository servers are VM's. So I shutdown and take a snapshot as my fall back point.
But this is Windows so...REBOOT! lol
Bill

[Mildur]

Hi Daniel

Our customer support found 3 similar cases and will check it out with our RnD Teams.

Best,
Fabian

[SamCOM]

Wished i'd browsed the forums yesterday prior to undertaking the V11 > V12 upgrade.

I also experienced the same issues as reported in this thread.
After failing the install, the V12 installer wanted all V12 Veeam B&R products uninstalled first. Going through uninstalling the failed V12 products and then running again worked using the existing databases.

Then took 2 hours to update the VALRedist.msi on the B&R server once opening the console.

All jobs are still there are still work.

[tivruski]

I had a similar experience recently. My installation took around 16 hours to complete (I left it running overnight), but I eventually aborted the process. I uninstalled everything and then installed V12 along with a new MS SQL database. Currently, I'm attempting to restore the configuration from backup. As of now, the process is stuck at "Upgrading database (78% done)", after 15 minutes have passed. SQL seems to be utilizing approximately 26% of the processor time, while other metrics, such as Database I/O, are at 0%. I plan to leave it running overnight and check the progress tomorrow, although I don't have much hope for success. Since I'm using the Community Edition and have only five VMs, I'm considering rebuilding everything from scratch if necessary.

[Mildur]

Hello

Our support team found a possible bug in MSSQL server 2012 (may be not limited to server 2012) that was affecting our installation process. The suggestion was to upgrade the MSSQL server used for the configuration database to the latest SP version or one of the newer MSSQL servers like 2019 and then try again.

If you have a case open, please ask you support engineer to check your case against the case from Daniel:
post480508.html#p480508

Best,
Fabian

[darryl]

BUMP so maybe more people will be aware that this still happens.

Looks like we ran into this one as well. Buried down in the thread list so I didn't see it before the upgrade.

Opened case 06032317 to see if there's an easy way out of this. Don't really want to be in a restore situation.

[masonit]

Had the same problem. After upgrading sql server from 2012 to 2019 it went fine.

[importantchicken562]

Hi,

Sorry to bump an old thread but I just had the same issue and it was disastrous. I had to uninstall everything, install SQL2016 express, install BR12, then import my config backup and revalidate all of my credentials to all of my VMS/repositories.

I need to upgrade 6 other Veeam backup servers. Since someone mentioned it is a bug related to SQL2012, would stopping all Veeam services, upgrading the SQL instance Veeam uses to 2016, starting the Veeam services, then running the upgrade work?


Thanks!

[wt0]

Results:
Removed all Veeam and reinstalled 11a with latest patch. New install created a new SQL instance (2016)
Restored config backup to get server back to where it was.
All backup jobs ran fine.
Started upgrade again.
All went fine this time. Upgrade took approx 20 minutes.

Happy camper

Will look at migrating to PostgreSQL in a couple days.

I'm trying to do the same. I'm back on v11 with clean backups. The enterprise manager wont re-install saying my existing DB is marked with a newer version than that which is installed. Changing the version then using the DB gave me missing patch errors.

[Mildur]

I'm trying to do the same. I'm back on v11 with clean backups. The enterprise manager wont re-install saying my existing DB is marked with a newer version than that which is installed. Changing the version then using the DB gave me missing patch errors.

We don't support a downgrade. You can keep Enterprise manager on V12 to manage a VBR v11a server.

Best,
Fabian