Question

What are the actual limitations of storage QoS?

  • 19 October 2020
  • 2 replies
  • 607 views

Userlevel 2
Badge +2

Hi!

While reading the documentation for storage QoS the following is mentioned as a limitation:

Storage QoS is not supported for Metro Availability, synchronous replication, and AFS virtual machines.

What will the actual result be if I apply storage QoS to a VM that I also replicate synchronously? Can I even do this? Or vice versa, what happens if I start replication of a VM with storage QoS?

In what way does storage QoS conflict with Metro functionality?

Thanks.


This topic has been closed for comments

2 replies

Userlevel 2
Badge +2

Thanks for the reply.

In this case we’re planning to use AHV, I forgot to mention that in my OP though.

None the less, the use of Storage QoS is limited so it shouldn’t be an issue, I just wanted to understand more about the limitations - and the official documentation needs to be vastly expanded upon in this regard.

Userlevel 6
Badge +5

Hi Martin,

 

This is an interesting question.
My initial thought was that the QoS adds to the cluster latency. The latter is crucial for Metro Availability as the commit of the data write happens on both sites. Increased latency in this instance could impact MA negatively.

Digging a bit deeper, it turns out that storage QoS is ingnored by the ESXi, at least the way the feature is implemented currently.