I'm trying to setup NSX-T in my homelab environment but am having issues on my Optiplex 7040 Micros.
This problem started from day one before I had began to setup any transport nodes or configure NSX.
I'm finding that the NSX manager VMs are consuming at MINIMUM 100% of a single CPU core via the rngd process and will cause nodes that usually sit around 20% CPU idle to idle at 60%+ CPU
What is happening is that the rngd process is consuming anywhere from 80-99% CPU (A single core) on any given host.
This has been especially noticeable on my Higher Clock lower core count systems.
The MHZ Consumption of the VM drops substantially (and in scale) when I move from high clock low core count to a high core count low clock CPU.
On a I5-6600t NSXMGR was consuming ~4ghz idle
On a E5-2650L V3 NSXMGR was consuming ~2ghz idle
I've tried lowering the CPU reservation in the VM's properties from 2000MHZ but that's made no difference.
If I kill the offending process rngd the CPU usage immediately plummets but the MGR will break after a couple of minutes.
I'm currently running a single controller deployment as NSXMGR is placing too much strain on my other hosts.
Currently running 18.104.22.168.0.19232594 underneath vCenter and ESXI as that's what's currently available via VMUG advantage.