What’s your automation setup for multi-node OpenNebula operations?

Solutions Hybrid Multi-Cloud
1

We’re managing a multi-site OpenNebula deployment and want to better automate VM provisioning, patching, and failover checks across nodes.

Curious what your tool stack looks like. Using native hooks/scripts? Ansible? Any experience integrating tools like Terraform or something more orchestration-focused?

Would love to learn from how others handle this.

2

Hello,

The best part of OpenNebula is that you have plenty of options for it, in my opinion (some of my colleagues may differ), when thinking about operations:

  • If you are used to work with terraform, the OpenNebula provider is the best way. Please check the docs at Terraform Registry for examples on its usage
  • The CLI allows everything to be done on a shell basis, but people differs about thinking if bash/zsh/mksh are or not programming languages. Thus, you have python, ruby, go and a XML RPC based API entry point (if your language has a libcurl and you can cope with XML it can be called)
  • About failover between checks, hooks that can be executed after certain API calls or in certain events. Please, note that some physical situations cannot be fixed (i.e., in case of a very sudden electric cut due to a fire, flooding, etc. the VM memory on the running host cannot be sent in time to another one)
3

We’ve built a hybrid setup that’s been working well for managing OpenNebula across nodes:

Ansible: for basic config mgmt (host setup, file sync, network settings)

Attune: really helpful for multi-step jobs like post-deploy validation, Ceph checks, snapshot cleanup, etc. You can run scripts across all nodes and resume if one fails

Terraform: for VM definitions + templating (especially for non-OpenNebula resources)

Bash + Cron: still using these for fast, non-critical routines

For monitoring: Prometheus + custom alerts via Alertmanager