Servers include the capability for IT service providers to install applications in Custom Virtual Machines (VMs). While the IT service provider is responsible for the management and maintenance of both the Custom VM’s operating system and application, business continuity can be assured by the server’s local backup processes and the optional Cloud Backup and Disaster Recovery product.
The Backup Process
Virtual Machine Volumes
Windows Server based Custom VMs include both a system volume (the C: drive) and a data volume (the D: drive).
Linux based Custom VMs include a system volume (\config), a data volume (\data), and a backup staging volume (\backup).
Each volume has a specific role:
- The system volume contains the operating system (OS) and the program files of any installed custom applications.
- The application must be configured to use the data volume to contain its working data files.
- The backup staging volume of Linux VMs is used to prepare data before it is backed up by the server.
Windows and Linux Custom VMs share a number of similarities in their backup processes:
- Backup starts daily at approximately 11pm.
- During the backup process files are backed-up, compressed and encrypted to a separate volume, not accessible to users, dedicated to local backups. Here, files are transported over the network to the optional Cloud Backup target.
- If the server includes the optional Cloud Backup and Disaster Recovery option, data in the local backup is streamed up to the cloud for long term storage. Once secured in the cloud, the local backup may be overwritten by other backup processes.
Windows Server Virtual Machine Backup Process
Servers will automatically employ a Windows Server VM application-consistent backup process for the VM’s system and data volumes.
The server will automatically request VM applications quiesce (pause their processes to ensure their live files are safe to backup) using the Windows Server VSS service. The server’s automated backup process then takes a snapshot of the system and data volumes. The snapshot process is almost instantaneous, allowing the application to resume with negligible impact on end users.
For applications that are not VSS-aware, and therefore cannot be automatically quiesced, a crash-consistent backup process is employed by default. The VM remains up and running during this time, with no effect on end users. For applications that are not VSS-aware but still require application-consistent backups, an alternative process can be used where the VM is temporarily shut-down out-of-hours before the backup takes place. For each VM requiring it, a request should be made to the Support team to configure this alternative backup process.
Linux Virtual Machine Backup Process
Servers will automatically backup data that has been copied to the VMs backup volume.
This allows two approaches to application backup to be implemented by an IT service provider:
Firstly, if supported by the application in the VM, the IT service provider should configure the application to schedule an export of its data to the backup volume at a suitable time before the nightly server backup takes place. This will provide application-consistent data to backup.
Alternatively, the IT service provider should configure the VM to schedule a simple copy of files from the VM’s data volume to the backup volume at a suitable time before the nightly server backup takes place. This provides crash-consistent data for the backup.
One or other of these approaches must be configured by the IT service provider to ensure the application’s data is correctly backed-up.
Local Backup Retention
With Cloud Backup
For servers that include the Cloud Backup and Disaster Recovery product option, data is backed-up locally only until the data has been successfully uploaded to the cloud.
The local backup area is able to hold data from several days worth of backup processes if necessary. So even in the event of a WAN outage or cloud storage issue that may cause the daily Cloud backup to fail, data will be regularly backed-up locally.
Without Cloud Backup
For servers that do not include the Cloud Backup and Disaster Recovery product option, the local backup area is sized so that it can hold enough historic data (a combination of a full VM backup and regular incremental changes) to ensure the VM can be restored back to its previous state when the last good backup was taken, in the event of a major loss of data.
Cloud Backup Retention
Custom VM files are backed-up to the cloud for servers that include the Cloud Backup and Disaster Recovery product option
Cloud storage is automatically sized so that it can hold enough historic data (a combination of a full VM backup and regular incremental changes) to ensure the VM can be restored back to its previous state when the last good backup was taken, in the event of a major loss of data.
In order to provide protection against events that may have resulted in corrupted data being backed-up to the cloud, such as a CryptoLocker infection or accidental deletion, thirty days data retention in the cloud is supported as part of the Cloud Backup and Disaster Recovery service.