When you back up a VM that runs an application or a database, such as Microsoft Exchange, Active Directory, SQL Server and others, it is critical to ensure that the data inside those application and database backups is consistent. This is important because portions of data can be kept in memory and some transactions may be incomplete at a time of a backup. If no steps are done to flush memory and I/O operations, such a backup would be crash-consistent, that is, an equivalent of pulling the server plug and then powering it back on. Of course, most modern applications and databases provide ways to recover from such a state but in most cases, you would still need to do some manual restore operations and could still run a risk of losing data.
To ensure application and database data consistency, NAKIVO Backup & Replication provides the Application-aware backup and replication feature, which is enabled by default for all backup and replication jobs. This means that all your backups and replicas can contain consistent application and database data, without performing any extra configuration steps. As a result, you can not only instantly recover VMs, which can contain consistent application and database data, but can also instantly recover Microsoft Exchange objects, such as emails, and recover Active Directory objects, such as users, directly from compressed and deduplicated backup repository.
For Windows-based VMs, NAKIVO Backup & Replication relies on VMware quiescing technology to perform application-consistent VM backup and replication. When a backup or replication job is started, NAKIVO Backup & Replication requests VMware vSphere to create a quiesced snapshot of a VM. Prior to making the snapshot, VMware vSphere triggers the Microsoft Volume Shadow Copy (VSS) service in the VM via VMware Tools. The VSS service instructs VSS-aware applications and databases to flush memory and I/O operations and then informs VMware vSphere that a snapshot can be taken. VMware vSphere then takes a VM snapshot in which application and database data is consistent.
During the snapshot creation process, applications and databases continue to run and save all changes to memory, so no downtime occurs. Once the snapshot is created, VMware vSphere triggers the VSS service again, which now informs applications and databases that they may flush data from memory to disk and continue normal operation. Finally, NAKIVO Backup & Replication reads data from the snapshot and sends it to a backup repository or to a target DR server. Once completed, NAKIVO Backup & Replication automatically removes the snapshot. As a result, backup data is consistent, while applications and databases are fully functional during backup.
What if your application is not VSS-aware or runs on Linux? NAKIVO Backup & Replication provides the ability to run custom pre-freeze and post-thaw scripts to enable application awareness. A pre-freeze script is executed before a snapshot of a VM is taken, and a post-thaw script is executed after the snapshot has been taken.