Overview

The following are best practices recommended for use when Cloud Connect Backup is being utilized.

Configure Weekly Retention on Jobs

With Backup & Replication version 11, new backup copy job GFS logic was introduced. With this change, jobs will now format backup chains using Forward Incremental Backup instead of Forever Forward Incremental Backup. In cases where monthly backups are configured, this can lead to an exponential increase in consumption due to the additional space needed for extra incremental files. For this reason, we recommend configuring at least one weekly restore point in job settings. This affects all jobs upgraded from version 10 and below, as well as new jobs. You can read about this change here.

Do Not Use Low-End NAS Devices as Repositories

While it is possible to write backups to a NAS device, this is constantly discouraged by Veeam employees on the R&D Forums. Here is one such example. When a NAS device is used, an iSCSI presentation of the storage is recommended. Instead, Veeam recommends the use of a server with hardware-backed RAID and battery-backed write cache to maximize performance and minimize corruption. Additionally, this allows for the use of ReFS or XFS, which increases merge speed, decreases disk space used for GFS points, and in the case of XFS, allows for immutable on-prem backups.

Do Not Use Deduplication Appliances for Primary Retention

While a deduplication appliance can be used to house the entire backup chain on-premises, it is not recommended as it will cause decreased backup and restore performance. Instead, it is recommended that an alternative target be used for short-term primary backups be utilized, with the deduplication appliance storing only long-term retention points. This short-term repository can be all-flash, hybrid-flash, or hardware-backed RAID managing spinning disks. Additionally, a special configuration may be needed for your deduplication appliance to function properly with Veeam. For more information, read this Veeam Knowledge Base article.

Utilize Backup Copy Jobs Instead of Backup Jobs

While regular backup jobs can target Cloud Connect repositories, it is recommended that backup copy jobs be utilized instead. When writing a backup job to cloud connect, production systems will be impacted for the duration of the job run. When utilizing backup copy jobs, backups are synthesized from existing backup files, reducing strain on production systems. In addition, backup copy jobs have much better logic for retrying failed sessions and the resumption of data transmission after an error. Finally, backup copy jobs can also leverage WAN acceleration with a Cloud Connect repository to reduce processing time. This functionality does not exist for regular backup jobs.