1. Cross-site DR does not support recovery of data stored in the physical disks, shared virtual disks, USB devices, and CDs/DVDs of VMs.

2. The primary/secondary DR function of SCP creates local BPs and packages and replicates backup data to the secondary site. When the primary site is down, the secondary site can directly recover VMs based on the latest backup data stored on the secondary site and ensure crash-consistency if the DR mode is used. In scheduled recovery mode, the latest data is first synced from the primary site to the secondary site.

DR policies are classified into the following consistency levels: Inconsistency backup, crash-consistency backup, and app-consistency backup. Inconsistency backup means that file or disk data is not captured at the same point in time. Crash-consistency backup means that disk data is captured at the same point in time, but in-memory data or silent app systems are not backed up. App-consistency backup means that disk and in-memory data are backed up simultaneously to ensure the consistency of app states. In crash-consistency scenarios, database instances may fail to be recovered from backups. This is because the instance data is still stored in memory. A log rollback operation is required to ensure the consistency of transactions.

3. If you have a database and require an RPO of 0, you need to use the database to implement DR or backup.