Troubleshooting OLVM backup and restore
If an oVirt checkpoint recorded in the NBU catalog is deleted or missing on the OLVM engine (orphaned / reset / manual virsh / oVirt admin action), the next scheduled incremental will automatically fall back to a full backup to avoid a chain gap.
Do not manually delete oVirt checkpoints for backed-up VMs outside NetBackup. NetBackup will manage the checkpoint chain.
If a previous OLVM backup was interrupted, some backup activity may still be active on the OLVM engine when the next job starts. This can cause conflicts, but NetBackup detects and resolves them automatically.
NetBackup handles these situations as follows:
If a data transfer from a previous backup is still in progress, NetBackup reuses it instead of starting a duplicate.
Before completing a backup, NetBackup clears any leftover transfers so that finalizing the backup does not fail.
When a new job starts, NetBackup completes any backups that were left unfinished by earlier runs, while preserving incremental (checkpoint) progress so that data does not need to be re-transferred.
Recommended action: If the engine still holds a stale backup, manually finalize the backup on the engine, then re-run the job.
If the async pipeline cannot start (thread creation/resource issue), OLVM transparently falls back to synchronous ImageIO I/O. Mid-transfer retry exhaustion aborts the disk and returns a retryable job error.
Recommended action: Re-run the job. Tune the OLVM_ASYNC_* bp.conf keys.
When you enable the option during a full VM recovery, NetBackup checks for and removes conflicting VMs before restoring, based on two criteria:
Matching display name and datacenter:
If a VM with the same display name already exists in the same datacenter, NetBackup overwrites it. VMs with the same name in other datacenters are not affected.
Matching UUID
If you also choose to restore the VM with its original UUID, and a VM with that same UUID already exists on the OLVM server, NetBackup removes the existing VM so the recovered VM can retain its original UUID.
If the option is disabled, NetBackup does not delete any existing VM, and the recovery fails if a conflicting VM is found.
To configure a KVM backup host, add a KVM_PROXY_SERVER entry in bp.conf/registry on the primary server for each NetBackup client you want to use.
For example:
KVM_PROXY_SERVER = nbclient1.com KVM_PROXY_SERVER = nbclient2.com
After this configuration is completed, NetBackup clients become available for selection as backup hosts in KVM policies.
NetBackup incremental backups of OLVM virtual machines may automatically enable the setting on VM disks, overriding an administrator's decision to leave the setting disabled in the OLVM/oVirt UI.
Recommended action: If you require to remain disabled for a disk, manually clear the checkbox in the OLVM/oVirt admin UI. Navigate to , clear the checkbox after each backup completes.