Becomes Google Open Source Spanner

Database: Google Cloud Spanner introduces backup / restore

Google has added a backup / restore function to its Cloud Spanner database service. The Database as a Service (DBaaS) is geared towards reliability and high availability, but the innovation is intended to cushion errors by users or by applications. The Internet giant has also added a local emulator to Cloud Spanner, and the C ++ Client Library is now generally available.

Backup / restore despite failure safety

The concept of Google Cloud Spanner is designed for high availability: It scales horizontally and, according to the promise for the official start of the service three years ago, should offer an availability of 99.999 percent. It also relies strictly on ACID transactions (Atomicity, Consistency, Isolation and Durability). An open source project based on the foundations of Cloud Spanner emphasizes the longevity of the concept with its name: CockroachDB has been crawling as a database "cockroach" in the cloud as a competitor to Google's offer since October 2019.

The backup / restore functions that have now been introduced serve to protect against operator and application errors that lead to data loss. The service encrypts the backups and keeps them available for up to a year after creation. Of course, there are additional costs for the backed up data. Users set an expiration date for the backup, on which the service deletes it. For longer backup copies, Google recommends exporting the database.

The copy moves with the original

The backups reside in the same instance as the source database, and Cloud Spanner always replicates them to other geographic locations along with the actual database. The backup copy contains all data including the schemas and secondary indexes of the original database for the given one.

The restore only works in a new target database, not in the original source. The new database must be in the same project as the backup and configured for the same geographic instance. The target of the restore retains all original data including schemas, but administrators must configure the access permissions separately.

Local playground

In addition to the backup / restore function, Google Cloud has expanded Spanner to include a local emulator, which currently still has beta status. It runs offline and emulates the Spanner API both as a REST interface and via gRPC and SQL access. Developers and administrators can use it to test the functions and data locally before they go live, which can cause unnecessary costs and problems.

A C ++ client library is now available for high-performance access to the data in Cloud Spanner. The library offers session pool management and retry logic. Further innovations such as the integration of foreign keys can be found on the Google Cloud blog. (rme)

Read comments (3) Go to homepage
Ad ad