HomeLab: nested vSAN Cluster (I/O failures)

[benj95]

I wanted to set up a nested vSAN cluster on my single ESXi host as base test environment for a NSX environment and in the first moment the nested Cluster works great, but after 1-2 days some disks in the vSAN become a I/O failure and some of the virtual machine will become inaccessible.

I’ve tested the scenario with different virtual machine settings and various vSphere versions (vSphere 6.7 and 7.0), even on another physical ESXi Host, but I have always the same result with the failure and the inaccessible virtual machine after a few days.

I have four virtual ESXI for the vSAN Cluster each with six vCPU, 16 GB ram, one network interface and four Disks (1x 32GB -> ESXI System, 1x 60GB -> vSAN Cache, 2x 260GB -> vSAN Capacity):

The only workaround that has helped is to regularly put the host into maintenance mode and restart it.

I know that the nested ESXI isn’t supported officially, but does anyone have the same problem and a solution for it or generally an idea what could help?

[depping · Accepted solution]

I never seen the issue, and I have virtual labs running for weeks sometimes. It wouldn't surprise me though if you are running out of memory for vSAN as well by the way, 16GB is not a lot to run the hypervisor and vSAN on top of that.

[benj95]

It was the ram from the host, with 24GB the vSAN works now since 3 Weeks perfectly.

Thank you for the solution!

[depping]

thanks for the feedback.