- This topic has 3 replies, 4 voices, and was last updated June 22, 2015 by Matthew C.
Performance Overhead
You must be logged in to create new topics.
Click here to login
Is there any documentation the discusses performance overhead (CPU/Write Latency….) when protecting a VM?
Zerto doesn’t impact protected VMs from a resource (CPU, RAM) or disk (I\O, throughput) perspective. Instead, Zerto adds a negligible CPU overhead to each ESX\ESXi host of typically 2-4%.
Any other overhead you’re considering?
There is some IO overhead in the VRAs also, but again it’s not a lot, if you run ESXi off local spinning disk you can store the VRA’s on there but it’s not required.
Like Ian said, I do run my VRA’s off of the local storage on each ESXi Host. They can’t move to another host anyway, so there is no need to burn up shared SAN storage or IO with them. They are so light I wouldn’t hesitate to put them on shared storage, but I’ve sort of made it a personal rule of thumb…any appliance that is fixed to a static host (like the VRA) runs off of local storage. If the host is unavailable by rule the VM is as well, 100% of time, so no reason not to run it locally. Just a personal preference but figured I’d share. I do the same thing for other appliances that run the same model, 1 agent VM per host and it can’t vmotion off.