Most efficient option for backup repository

[ssandoz]

Hi,

I currently run many backup jobs in VEEAM B&R, all pointing on a single backup repository.
Due to a lack of resources for the server defined as backup repository, I can only run 6 concurrent tasks.

Now I will have the opportunity to rebuild a new storage infrastructure to store backup files and I wonder what should be the most efficient solution:

Case 1 - Directly attach a big storage to the VEEAM server. In that case, should I need to increase resources for the VEEAM server as it will be used as backup proxy and backup repository?
Case 2 - Keep the same configuration as now. That means a single server with a big storage, referenced as backup repository. But this time, the server will have much more resources to allow more concurrent tasks.
Case 3 - Split my storage volume. That means created multiple servers with smaller storage and reference all of them as backup repositories. Then change job settings to point on one of these backup repositories.

Do you have any other solution I did not think about?

Thanks in advance for your support.

[HannesK]

Hello,
without more details (amount of data, number of VMs, backup window) I can only give generic advise.

If you only change storage, then you can sill only run the same amount of tasks. That's because you need to increase CPU / RAM for more concurrent tasks.

Case1: is used for customers that want redundant storage controllers. A good option with FibreChannel / iSCSI / (SAS), yes.
Case2: Up to 1000 VMs is my rule of thumb for a high density server (HPE Apollo, Cisco S3260 and others). A high density server is probably the best price / performance product today
Case3: sure, you can also go with smaller servers. Scale-Out repositories might simplify the setup.

Best regards,
Hannes

[ssandoz]

Hi,

Thanks a lot for your feedback.
At the moment, we are speaking about 31 backup jobs
- 17 jobs managing 17 physical server backup (1 per job) for a total amount of data of 120TB
- 14 jobs managing 71 VMs for a total amount of data of 63TB
Backup window is during the night so from 10PM to 7AM.

But more backup will come later so that's why I will like to select the best option.

[HannesK]

customers have a different opinion on what is "best". All options above are fine for your scenario.

For simplicity I would go with the "all-in-one" server. KISS - https://en.wikipedia.org/wiki/KISS_principle