@LTS_Tom, on this week’s Thursday VLOG you and Jason both commented about
problems with Azure HCI as a hypervisor in relation to Vmware and XCP.
This is a common critique I have heard from those in the non-Microsoft camp but I was
wondering if you have any specific resources or references you can point us to that could
help further elucidate the shortcomings/pitfalls of this platform.
I totally appreciate and respect the non-MS camp and there are many valid reasons for this philosophically, but what’s hard for me is that the MS MVP’s all talk about Azure HCI as a perfect
solution and the non-MS people all say it’s problematic.
What can a poor IT procurement person do to make a better decision?
I will see if Jason has time to add to this, but the short answer would be just how buggy the system is especially around the updates. The primary complaint Jason brought up is that it still has Windows as the base OS which is what both causes the update issues and makes it much more bloated.
What architecture are you looking to use for your buildout? If using converged or hyper-converged I would expect some to be better than others which may not be the same ones when compared to a setup using local resources.
Ideally we are looking at the SuperMicro big twins either in the 2U or 4U flavor.
That gives us a combined 24 disks across the nodes and depending on which hypervisor
we use a lot of resiliency options (Vmware Vsan Storage Policies, HyperV host resiliency config, etc).
You are correct the other approach would be to strip down the hosts to just compute and VM local storage and then add some kind of SAN type or NAS type shared storage but we are trying to be somewhat hardware frugal and we also are trying to create an easy to replicate stamp as this will eventually scale outwardly to numerous distant sites.
Ok, so I’m guessing you are looking at a 2 node design? Or would you have more than one box at a site?