control your cluster of virtual machines
Go to file
Jeff Carr 7a4bc0b5d6 track hypervisor daemon kill count
Signed-off-by: Jeff Carr <jcarr@wit.com>
2024-10-13 00:57:29 -05:00
resources the beginning of something 2024-10-12 00:17:26 -05:00
.gitignore initial something. doesn't do anything yet. 2024-10-11 13:55:13 -05:00
LICENSE initial something. doesn't do anything yet. 2024-10-11 13:55:13 -05:00
Makefile the beginning of something 2024-10-12 00:17:26 -05:00
README watchdog for each hypervisor 2024-10-12 11:21:45 -05:00
argv.go add daemon mode 2024-10-12 19:44:43 -05:00
configfiles.go first basic check to tell if the cluster is healthy 2024-10-12 12:45:43 -05:00
event.go ugly but limping along progress 2024-10-12 10:59:11 -05:00
http.go track hypervisor daemon kill count 2024-10-13 00:57:29 -05:00
main.go cleanup STDOUT debugging 2024-10-13 00:40:22 -05:00
poll.go track hypervisor daemon kill count 2024-10-13 00:57:29 -05:00
structs.go track hypervisor daemon kill count 2024-10-13 00:57:29 -05:00
watchdog.go polling duration times work 2024-10-12 11:54:01 -05:00

README

virtigo: a control panel for your virtual machine cluster

This is an attempt to make something that should:

* Maintain the master list of virtual machines that should be running at all times
* Work with a cluster of dom0 hypervisiors via libvirt and/or qemu
* Understand the cluster and autostart droplets that have failed
* Initiate and Handle all droplet events: start, stop, clone, migrate, etc
* Verify consistancy on the hypervisors
* Initiate and handle live migrations between hypervisors
* Talk to the hypervisors using only the virtigo daemon via protobuf
* Work in GUI mode (GTK/QT/etc) but ALSO the console (ncurses)
* GPL'd with the intent for use with homelab and personal hobbyists

Notes & Goals:

* Be super easy to use.
* Automatically map access to serial and graphical consoles
* This is intended for managing Virtual Machines, not for containers
* This often uses the DO nomenclature 'droplets' instead of 'virtual machines' or 'domU'
* Every droplet is considered hostile
* When possible, use protobuf
* Let security be handled externally at the socket layer with other tools
* Put network, cpu, etc stats in external tools
* Automatic live migration to decommission nodes
* Implement iptable rules via the virtigo daemon

Inspired by:

* kvm
* virt-manager
* golang
* the simplicity of the DO droplet control panel