B&R Server restart killing Repository

[Hauke]

Case #04784828

Hello,

I currently testing to switch to Linux repositories. Everything worked great over some days, no issues, but after rebooting the B&R Server it's now marking the repository as "Outdated", Job failing:

Code: Select all

Failed to synchronize LinuxStorage1 Details: Product components installed on x.x.x.x have older version (11.0.0.837) than the current product version (11.0.0.837) and must be updated.

If I try to use the Upgrade-Function it's failing with:

Code: Select all

Failed to upgrade host components. Error: 'Invalid package type DeployerSvc'

I already removed everything from Veeam, added the Repository again, recreated the Backup Job - and it worked, until the next reboot. Same error.
Anyone had such issues in the past?


Veeam B&R: 11.0.0.837 P20210401
Repository: Ubuntu Linux amd64, 20.04.2

[Dima P.]

Hello Hauke,

Can you open the managed Linux server wizard for the mention repository and click it through the very end? There might component mismatch after upgrade and the component check, which is performed on the apply step of the wizard, should solve that. Let us know how it goes!

[Hauke]

Wizards are not doing anything, they are stopping with the same error message "Product components installed have older version (11.0.0.837) than the current product version" before they do anything. The only thing working is to completely remove the Jobs, Repository and Managed Server and add them again. Then the Wizards are running through without any error, I can even do the backups to the repository, but again after a reboot of the B&R Server I get the same error again.

[Dima P.]

Thanks for the update! Please keep working with our support team as the described behavior does not look expected. Let us know how the investigation goes or if you need any assistance with support team.

Гpdate from QA team: when you notice that repository is inactive after another reboot, сan you please try to Rescan all managed servers under Backup Infrastructure > Managed Servers (you need to right click the Managed Servers node and hit Rescan).

[Hauke]

Rescan all managed Servers is saying "rescan successfull" for that repository, but it is still listed unter "out of date".

But you were right with your first answer, I just found out that another wizard is starting when selecting "properties..." of the managed server from the repository - this wizard is running through and updating the host successfull, the repository is working again after that.
But: after a test restart of B&R server it is listed as outdated again.

For me it looks like B&R is running a job on startup (and maybe scheduled too?) to check the versions of the connected servers, and this job isn't getting it right?

[Dima P.]

Hauke,

Any chance Linux repository was somehow cloned from any other machine already registered in Veeam B&R console? Thanks!

[Hauke]

No clone, fresh install from Ubuntu .iso directly to the NAS device.
But that NAS was used with Windows before, so the MAC adresses are reused.

And the old managed server name of it installed with Windows is still existing in Veeam as disconnected because it isn't possible to delete it - it was used in tape backups jobs, this jobs are blocking the deletion and that jobs can't be delete too (as there is no possibility to delete old tape jobs).

[Dima P.]

Thanks for the update! QA team has requested Veeam B&R database to perform record mismatch checks, support folks should follow up shortly.

[Hauke]

Ticket was solved successfully.
Issue was caused by duplicate host IDs after reinstalling the NAS from Windows to Linux without deleting it first in Veeam.
Database needed to be fixed and the support provided me the exact guide what to do.
Everything working now as expected.