125 VMs - Best Practices?

[BCT-Tech]

Would like some suggestions on best practices for backing up this large environment:
VMware ESXi 5.5
vCenter installed on 2012 VM
Veeam installed on 2012 VM
2 Clusters
7 ESXi Hosts
14 Datastores on 2 SANs
All Hosts have access to all Datastores
125 VMs
NAS Backup Repository with 10 TB

Several of the VMs have different credentials for application-aware processing. We would like to be able to vMotion VMs between Hosts and not have to re-configure the backups.

1. How many jobs should we create?
2. Should the jobs run concurrently, or one after the other?
3. How many Backup Proxies should we have?
4. How many concurrent tasks should each Proxy have?
5. How do we get the best throughput?
6. Any other configurations that would help?

Thank you for your help!

[foggy]

Several of the VMs have different credentials for application-aware processing.

Not an issue, you can specify individual credentials for some VMs in the job settings.

We would like to be able to vMotion VMs between Hosts and not have to re-configure the backups.

If VMs are being constantly moved between hosts and to be sure newly added VMs are automatically picked up, you need to add VMs into jobs using containers, such as VM folders.

1. How many jobs should we create?

You can group VMs using any factor that is reasonable in your case. It can be grouping critical VMs together to have lower RPO, for example, or VMs that should have similar retention. Also keep in mind that Veeam B&R performs dedupe between VMs in the job, so the more VMs with similar OS (or even created from the same template) the job has, the higher space savings are.

2. Should the jobs run concurrently, or one after the other?

We recommend to schedule the jobs to run concurrently, set the acceptable concurrent tasks limits on proxies/repositories, and let Veeam B&R automatic resource scheduler do its job.

3. How many Backup Proxies should we have?

Only actual tests will show how many proxies you need to meet your required backup window with the backup load you are going to assign to them (keeping in mind the amount of changes occurred inside VMs, proxy servers specs, required RPO values, etc).

4. How many concurrent tasks should each Proxy have?

Depends on the proxy specs. Also keep in mind that each task requires a single CPU core.

5. How do we get the best throughput?

Decide on what transport mode you want to utilize, what backup method your target storage is capable of in terms of IOPS. Start backing up, analyze the bottlenecks and optimize the setup further.

6. Any other configurations that would help?

If your production SANs are among supported in terms of storage integration, consider using storage snapshots.