Design question - multi backups per day - secondary target

[geordi]

Hi

I am thinking/considering a re-design of our current Veeam setup, hopefully some of you can give me some hints for that.

Following situation:
8 ESX Hosts with local storage (DAS based on HDD and SSD)
1 Backup Proxy per Host
2 Backup Repositorys (physical Windows Servers using Dedup Feature)
Dedicated physical network for backup-traffic and connection
150 VMs running in 25 Jobs

What we would like to achieve:
Monday to Friday at least 4 backups during the day (eg. 8 am, 12am, 4pm, 8 pm)
Saturday one backup
Sunday one backup

The Mo-Fr backups should be at least be available for one week, all systems must have a minimum of one restore point every day for 32 days.

Current design:
At the moment we are doing incremental backups with ActiveFull on Saturday and 100 restore points completely managed/scheduled via Powershell-Script. Problem here is the waste of backup storage due to the high number of restore points resulting in up to 5 Active Full and roundabout 95 Incrementals.

My questions:
1. What would be the best design here for achieving our goals but not wasting backup space?

At the moment I am thinking of having a job with:
Monday to Friday Increment 4 Backups per day
Saturday Active Full
Sunday Increment
22 Restore Points

Afterwards a BackupCopy/Secondary Target Job which will hold 1 restore point per day for at least 32 days.
This will result (if I calculate correct) in roundabout 1 Active Full and 21 Increment for the primary Job and 4 weekly fulls and 28 Increments on the Secondary Target.

2. The rotation in Veeam is based on number of restore points .. is there any way to handle that in number of days we would like to keep or do we have to count the restore points to match the days we would like to have (as we are doing it right now)?

Regards,
geordi

[foggy]

Hi Michael, why not to switch the backup job to forever incremental mode without any active periodic fulls that take most of the space? As for your second question, there's currently no ability to specify the number of days to keep.

[geordi]

Hi foggy

My current knowledge is that periodical/weekly active full and inc is the best-practice when using Windows Dedup or generally when using Dedup-Appliances.
Has this being changed?

Regards,
Michael

[foggy]

This recommendation comes mostly from performance considerations (no merge operations with periodic active fulls). On dedupe storage full backups should be deduped anyway, so space shouldn't be your major concern then.

[geordi]

So my current idea with using a backup copy is not necessary you mean?

[foggy]

Backup copy would still be required for secondary backups with longer retention.

[EricDo]

Design question overhere too.
Multiple huge backups during all week.
During daytimes.
20 TB to a single disk station a Dr4100 from DELL Hypercluster and a SQL cluster are writing separate to that disk station. A copy to staging is running always. From that disk station a backup to tape full is starting on friday. SQL and file backups.
During the week the whole week are running incrementals to disk and to tape.
Resulting in backuups running all week without any pause.
Daytime and during the night.
I want to split things up because the schedule is to tight. A second disk station, and maybe a second tape library. But is it possible te have two VEEAM backup servers on one network?
This to really split up things..

[foggy]

Hi Eric, I'd look into abilities to optimize the backup window. Have you observed the jobs' bottleneck stats?