Here is my host hardware… it consists of three idendical systems.

• Intel NUC13ANHi7 (RNUC13ANHI7000U) NUC 13 Pro 13th Gen Core i7 (Arena Canyon) Barebone Kit
• Crucial RAM 64GB Kit (2x32GB) DDR4 3200MHz CL22
• WD_BLACK 2TB SN850X NVMe Internal Gaming SSD Solid State Drive - Gen4 PCIe, M.2 2280, Up to 7,300 MB/s
• Crucial BX500 480GB 3D NAND SATA 2.5-Inch Internal SSD, up to 540MB/s
• I also picked up some new ethernet cables and a 2.5GB switch

I created a bootable USB with the minimal Ubuntu Server. Powered on the system and installed Ubuntu from the USB to the SSD (not the NVME)

In hindsight there are these little grommets on the bottom of the case the coincides with the screw ports in the SSD drive. I’m meant to remove the grommet and use a screw in the hardware bag/ziplock.

I installed ethtool for a moment until I realized that my nodes were running at 2.5GB… So now…

some convenience setup…

• ssh-copy-id -i ~/.ssh/id_rsa USER@HOST
• copy my sudoers file

scp SUDOFILE HOST:
ssh HOST
sudo sh -c "cp SUDOFILE /etc/sudoers.d/. && chown root:root /etc/sudoers.d/SUDOFILE && chmod 0600 /etc/sudoers.d/SUDOFILE"

microcloud dependencies

• sudo apt install libvirt-daemon

begin the LXD setup

• time to plug into the network and UPS
• then install snap if not already installed
• then install snap install microcloud lxd microceph microovn the tools on each node
• then init microcloud init

When something goes wrong

I don’t really like my notes above. While they are simple when they fail I do not know how to restart the process without reinstalling the host OS. And maybe that’s why the demos start in a nested VM.

Reference:

• Intorducing MicroCloud

Notes:

One thing I have not addressed here… there are a couple of BIOS flags that should be set/cleared… like AMD-V and IOMMU on AMD processors. Intel might be VT-d/IOMMU.

commands

microceph.ceph osd tree microceph cluster list lxc cluster list lxc list lxc launch ubuntu:22.04