How to properly program jobs

[jamb]

Hi!
I have a setup right now in which I have multiple jobs, grouping VMs by OS and chained together so one of the jobs starts at a certain hour and the next one gets triggered upon completion of the previous one.
Now, I came across this article: https://www.veeam.com/blog/worst-backup-practices.html in which it's stated that

"The Veeam Backup & Replication scheduler runs more jobs simultaneously to optimize the backup time window by balancing the load on backup infrastructure components, choosing the transport mode and limiting bandwidth to the repository."

So, does this mean that I should schedule all my jobs for the same time and let Veeam manage the resources and how many simultaneous ones are running instead of linking all the jobs in a single chain?

Thanks!

[oleg.feoktistov]

Hi Jon,

I think, both approaches make sense in distinct situations.
If it is crucial for you to have your VMs/Applications processed at a separate time due to some business reasons etc., then job chaining is an option.
If you want to minimize your backup window (e.g. run your jobs out of business hours only) and have multiple backup components at your disposal, then why not go with the option 2?
How many jobs to run simultaneously would depend on Concurrent tasks settings you apply to your Backup Proxies and Repositories.

Recommendations would be:
- For Backup Proxy - 1 task per core/vCPU, where 1 task = 1 VM disk processing.
- For Backup Repository - 2GB per core/vCPU + 1 task per core/vCPU, where 1 task:

• For regular backup chains = 1 job processing.

• For per-VM backup chains = 1 VM backup chain processing.

Besides, consider setting Read/Write Data Rates for Backup Repository depending on repository capabilities and your needs.
Also, check our Best Practices for Proxy Servers page.

Best regards,
Oleg