Backup I/O control

[Cragdoo]

I've been looking for some more in depth information around Bakup I/O control. As I understand it , the storage latency is polled every 20s. My questions are , if when polled and the limits are breached, are the jobs throttled then? And lets say at the next poll, 20s later the limits are back to normal, will the job throttling be removed then? Or is this on a per VM basis ?

[Shestakov]

Hello Craig,
I/O control is described in the User guide.
There are 2 latency thresholds, when first is reached, VBR stops assigning new tasks to datastore and in a case of the second one, Veeam Transport Service working with this datastore slows down data retrieval or writing.
Once your latency is normal, these mechanisms will be disabled.
Thanks!

[Cragdoo]

Thanks Shestakov, I had read that, but just wanted clarification if the throttling is applied at each 20sec polling period.

[Shestakov]

You are welcome.
To be exact, tasks check "assigning new tasks to datastore" flag once in a minute, while "I/O of existing tasks throttling" checks latency every 20 seconds.
By the way, there was a session on VeeamON called ​"Healthier Virtual Infrastructure with Backup I/O Control" which made a deep-dive to the backup IO control.
You can contact Veeam Vanguard coordinator to get access to the recording.

[Cragdoo]

Got it ...thanks

[Shestakov]

You are welcome.
Once you have a question, let me know.