v8.7.124 — Virtualization · Remote Console · Monitoring

Run every VM on every host
— from one screen.

KVM · libvirt · Proxmox · FreeBSD bhyve · Windows Hyper-V — one agent, one control plane. The web console bhyve never had: create guests, manage disks, open noVNC or serial. Kernel-level metrics underneath via DTrace · eBPF · ETW, every second.

4
Hypervisors, one agent
< 5MB
Binary size
0 deps
Zero dependencies

Up and running in 3 minutes

FreeBSD — one command

Download, install, and register the service

curl -L https://cdn.mihosrc.com/releases/mihosrcd_freebsd_amd64_8.4.2.tar.gz \
  | tar -xzf - -C /tmp && cd /tmp/mihosrcd && sudo sh install.sh

Key Features

Real-time Metrics

CPU, memory, disk, and network collected once per second. Optimized for FreeBSD via sysctl · getifaddrs · devstat.

DTrace · eBPF · ETW Tracing

FreeBSD sched:::on-cpu / off-cpu · Linux kprobe·tracepoint · Windows ETW NT Kernel Logger. Kernel-level scheduler latency, I/O, and memory events across all three platforms.

Unified Virtualization

Manage KVM · libvirt · Proxmox · FreeBSD bhyve · Windows Hyper-V guests across every node on one screen. Create, delete, power-control, snapshot, back up, clone — plus multi-disk, storage pools, Ceph RBD, and UEFI/TPM.

In-browser Remote Console

noVNC 1.7.0 graphics and serial consoles (virsh console · pct enter · qm terminal) tunneled to the browser over libwebsockets — no public IP or open firewall port on the guest.

Guest Auto-provisioning

Detects the guest OS (Linux · FreeBSD · Windows) and configures a serial console on offline images. 8.7.124 fully rewrote the Linux systemd getty and GRUB path.

Multi-node & Wireless Routing

Many WireGuard-linked nodes act as one. Wireless-only nodes route to guests via a proxy-ARP bridge, and ISO upload/management uses URL-transparent storage.

8.7.124 · Virtualization

Beyond monitoring — manage the VMs themselves

Every node across KVM · libvirt · Proxmox on one screen. From guest lifecycle to storage and consoles, handled by mihosrcd alone.

Guest lifecycle

Create, delete, power-control guests and change name, memory, vCPU, boot order, and CD/DVD.

  • Create · delete · power (start/stop/reboot)
  • Online setmem · setvcpus
  • Boot order · CD/DVD media (change-media)
  • Domain rename (domrename)

Snapshot · backup · clone

Capture state, back up whole, and clone. Block-commit collapses snapshot chains.

  • Snapshot create · revert · delete
  • Full backup (blockcommit-based)
  • Guest clone
  • Per-disk handling on multi-disk

Storage · disks

Create storage pools and volumes, attach disks, and handle Ceph RBD and UEFI/TPM guests.

  • Directory storage pool · volume management
  • Multi-disk attach/detach
  • Ceph RBD backend
  • UEFI · TPM guest support

Remote console · provisioning

Open noVNC graphics and serial consoles over a relay tunnel, and auto-configure serial after OS detection.

  • noVNC 1.7.0 graphics console
  • Serial console (virsh console · pct · qm)
  • Guest OS auto-detect (Linux · FreeBSD · Windows)
  • Linux/FreeBSD serial console auto-setup
KVM · libvirtProxmox VEFreeBSD bhyveWindows Hyper-V
View virtualization
Cloud Build

From a single node to a private cloud

One step past monitoring and virtualization: pool nodes to run like a cloud, and move guests between them.

01

Pool nodes

Many WireGuard-linked hosts as one resource pool. See the inventory on one screen even when KVM, libvirt, and Proxmox are mixed.

02

Place capacity

Decide where a guest lands from per-node vCPU, memory, storage headroom, and overcommit ratios.

03

Move guests

Relocate guest OSes between nodes. Offline (stopped) migration first, with shared-storage migration planned.

04

Storage tiers

From directory pools to Ceph RBD shared storage. On a shared backend, migration skips the disk copy and gets faster.

Guest OS migration

Minimize downtime when moving a running service to another node.

Roadmap · Live (zero-downtime) migration is on the roadmap.

  • Offline move — stop, copy the disk image to the target node, redefine
  • Shared-storage move — on Ceph RBD, move only the definition, no disk copy
  • Post-move verification — boot, serial console, and network on the target
  • Rollback — keep the source guest intact if the move fails

Supported Platforms

🐡
FreeBSD 13.x / 14.x
DTrace
Primary
🐧
Linux (Debian/Ubuntu)
eBPF
🎩
Linux (RHEL/Rocky)
eBPF
🪟
Windows Server 2016+
ETW
🍎
macOS (Apple Silicon)
kperf

Get started right now

Free download · 3-minute setup · Instant monitoring

Enterprise

We can help you build it

For multi-node private cloud design, on-prem builds, or migration consulting, get in touch.

  • Multi-node, multi-hypervisor architecture design
  • Migration from an existing VM estate onto mihosrcd
  • Ceph and storage-tier design with capacity planning
  • On-prem installation and operations support