Enhanced incremental backups rely on local database files located in <DP_home>\enhincrdb\MountPoint, if there are shared mountpoints in a cluster this leads to full backup for respective objects whenever failover occurs.
To avoid this enhanced incremental database for respective mount point can be linked to it:
1. In the shared mount point create directory called enhincrrep (eg: E:\enhincrrep is used as an example in the commands)
2. On active node create junction point to it with
mklink /J DP_home\enhincrdb\MountPoint MountPoint\enhincrrep
eg: mklink /J C:\ProgramData\OmniBack\enhincrdb\E E:\enhincrrep
NOTE: it cannot ne created if it exists already, you will need to delete DP_home\enhincrdb\MountPoint first
3. Failover to passive node
4. Delete DP_home\enhincrdb\MountPoint folder (just the shared volume folder, so that we don’t affect other volumes databases)
5. mklink /J DP_home\enhincrdb\MountPoint MountPoint\enhincrrep
On *ux this can be done with ln -s MountPoint/enhincrrep /var/opt/omni/enhincrdb/MountPointDir
This will allow the incremental chain to continue even after failover.