Our environment is vSphere (with a mix of Windows 2008, 2012, and 2016). A B&R physical box with dedicated attached storage. Also a Cloud Connect server in Azure.
We run a nightly backup of all the vSphere to the B&R server. We are configured for 21 restore points, weekly synthetic fulls on Saturdays, and backup file health check on Fridays.
We run a daily backup copy job to Cloud Connect in Azure. The source for this copy job is the existing backups. 7 restore points are configured, with GFS (4 weekly, 3 monthly, 4 quarterly, and 2 yearly). We also have backup files health check set to run weekly on Fridays.
I have a few questions since we are somewhat new to Veeam.
Since we run synthetic fulls on Saturdays with a backup file health check on Fridays, do I still need to create active fulls periodically? If so why and how often? If not why?
When the backup copy job is run, how does it create a full backup file? Does it send the full over the WAN, or does it get synthetically created on the CC server?
Also, where is the backup file health check performed for the backup copy. Is it performed on the B&R locally, or the CC server in Azure?
Is my configuration sound or do you recommend any changes?
F182 wrote:Since we run synthetic fulls on Saturdays with a backup file health check on Fridays, do I still need to create active fulls periodically? If so why and how often? If not why?
If the current approach fits your backup strategy, then Active full is not required. Basically, you are the one to decide how many full backups you would like to keep and how they should be created: synthesized from incremental restore point or active full from production.
F182 wrote:When the backup copy job is run, how does it create a full backup file? Does it send the full over the WAN, or does it get synthetically created on the CC server?
Though, for Synthetic full vs Active full, is there a risk difference? For our environment, we would rather not do periodic active fulls since some of our servers may take more than a day to run a full backup. We would rather run synthetic fulls moving forward after the original first active full (when the job was created). Though, we want to ensure the backups are good and without corruption. We need to be very sure we can restore without issue. This is why I question the "active" full option in the job configuration. Is this something I need to run once per month or once per every 3 months. Etc.
There is no difference between active full or synthetic full when you look at it from health check angle, so you are good to go with synthetic full backup option (makes perfect sense since you want to offload the pressure from your production storage).