After moving the vSphere infrastructure management to another vCenter Server, the Veeam Cloud Connect backup copy job fails showing the error "Unable to convert host Id...".
Analyzing the backup copy job error, the system reports the following issue:
Error: Unable to convert host Id from service provider database to host Id from local database: cannot find local host
The error is related to the Managed Object Reference ID (MoRef) that changes when the vCenter is moved or removed and re-added to the inventory. MoRef ID is a unique identifier used by vSphere to track objects such as datacenters, clusters, hosts, datastores, VMs, vSwitches and so on.
Veeam tracks virtual machines in its jobs through the MoRef ID but Cloud Connect jobs currently don't support the redeployment of the source infrastructure. Veeam performs a mapping between entities in Tenant's and SP's database and if the source infrastructure changes (new backup objects are created, which are mapped on the new hosts) the mapping breaks as well.
After struggling to find a fix, the Veeam support suggested me the following procedure that solved the issue.
Re-mapping the backup chain
From the Tenant side disable the failed Cloud Connect backup copy job.
The Provider side has to rename the Tenant .vbm file to .vbm_old.
From Tenant side, access the Backup Infrastructure area and right click the cloud repository. Select Rescan option to rescan the repository.
The Cloud Connect repository scan is being performed.
Now tell the Provider to rename the .vbm_old back to .vbm.
From Tenant side rescan the Cloud Connect repository once again.
Go to Backup & Replication area, right click the Cloud Connect backup copy job and select Edit.
Go to Target tab and click on Map backup link to map back to the existing chain.
Expand Backup Repositories and select the correct backup job. Click OK to map the selected backup chain.
Enable the Cloud Connect backup copy job. Right click the job and select Disable to remove the flag.
The backup copy job starts processing the backup.
After completing the remapping procedure, the backup job is now executed without any problem. The (Incremental) item after the backup job's name tells you the job is performing an incremental backup type, possible only if the existing backup chain (which it means at least a Full backup is present) is correctly mapped.
Depending on the available bandwidth, the job can take several minutes to complete.
If the "Unable to convert host Id..." error raises again, the fix is pretty simple and quick.