This page describes how to address most frequent inquiries about backup and restore issues.
Yes, in AlloyDB, both on-demand and automated backups have a degree of independence from the original cluster. This means you can restore data from these backups even after the cluster that created them is deleted.
Deleting a cluster does not prevent you from using the backup data to restore your data to a new AlloyDB cluster.
AlloyDB backup retention periods vary by backup type:
- Continuous: Retention is configurable from 1 to 35 days.
- Automated and on-demand: Retention can be up to one year.
Point-in-time recovery (PITR) in AlloyDB lets you restore a cluster to its state at any moment within a defined period. However, PITR has these requirements:
-
Cluster existence: the cluster you want to restore must exist. You cannot use PITR to recover a deleted cluster.
-
Continuous backup enabled: PITR relies on the continuous backup and recovery feature in AlloyDB. You can only restore to a point in time after the oldest continuous backup was taken. This means that if continuous backup was recently enabled, you might not be able to restore to a time before that point.
To restore your AlloyDB cluster to a specific point in time, see Perform point-in-time recovery .
Backups can only be restored within the same region as the backup. However, if you need to restore the cluster into a different region, you must first perform an on-demand backup of the original cluster, ensuring this backup is created in the target region. Then, you use that specific on-demand backup to complete the restore process in the new region. For more information, see Restore a cross-regional backup .
AlloyDB backup storage locations are determined as follows:
- Continuous backups are stored in the region of the cluster.
- Automated backups are stored in the region of the cluster by default.
- On-demand backups are stored in the region of the cluster by default but can be stored in a different region.
