Adventures of a Small Time OpenStack Sysadmin Chapter 040 - Nova Compute Service

Adventures of a Small Time OpenStack Sysadmin relate the experience of converting a small VMware cluster into two small OpenStack clusters, and the adventures and friends I made along the way.

Adventures of a Small Time OpenStack Sysadmin Chapter 040 - Nova Compute Service

First, links to some reference docs I used:

Nova Docs Page

https://docs.openstack.org/nova/yoga/

Kolla-Ansible Deployment Configuration Reference for Nova

https://docs.openstack.org/kolla-ansible/yoga/reference/compute/nova-guide.html

Nova Admin Guide

https://docs.openstack.org/nova/yoga/admin/index.html

Nova API Guide

https://docs.openstack.org/api-guide/compute/

Nova API Reference

https://docs.openstack.org/api-ref/compute/

Install CLI client

The CLI add on library for Nova should be installed by default.  Regardless, this link:

https://gitlab.com/SpringCitySolutionsLLC/openstack-scripts/-/blob/master/installcli/installcli.sh

Installs all the CLI addons I'm using, or just run:

 pip install python-novaclient -c https://releases.openstack.org/constraints/upper/yoga

Serial Port emulation for console and console logging

The console/serial port situation in Kolla-Ansible roughly matches the Install tutorial configuration on cluster1, which works well.  Its highly configurable but I did not need to change anything from default.  Cool!

Quotas

https://docs.openstack.org/nova/yoga/user/quotas.html

I set my Nova quotas in my project Heat Orchestration Templates such as this example:

https://gitlab.com/SpringCitySolutionsLLC/openstack-scripts/-/blob/master/projects/iot/iot/iot.yml

This Heat resource paragraph for the iot project shows specifically how its done:

  nova_quota:
    type: OS::Nova::Quota
    properties:
      project: iot
      cores: -1
      instances: -1
      key_pairs: -1
      metadata_items: -1
      ram: -1
      server_group_members: -1
      server_groups: -1

You can see I'm not all into setting quotas on myself.  Not a useful feature for my system requirements.  But it seems to work and is easy to configure (or configure to essentially ignore)

Usage Stats demo script

https://docs.openstack.org/nova/yoga/admin/common/nova-show-usage-statistics-for-hosts-instances.html#show-instance-usage-statistics

See this link for a demo script:

https://gitlab.com/SpringCitySolutionsLLC/openstack-scripts/-/tree/master/demos/nova

This outputs a table of project name vs servers configured, RAM MB-hrs, CPU-hrs, and Disk-hrs used.  Cool, easy to use.

Flavor Loader

I have written a flavor loader script, which is part of openstack-scripts project:

https://gitlab.com/SpringCitySolutionsLLC/openstack-scripts/-/blob/master/flavors/flavors.sh

I like my own flavor name format.  I know some, or many, people really like AWS style flavor names, but I don't need compatibility with that format and find my naming strategy very logical and easy to use.  I wish OpenStack did not use flavors; the way VMware lets an admin type in anything they want that fits within quota is so much simpler to use and more versatile.

Tomorrow it's not Mon-Day its Cinder-Day.  I really should have posted that on a Sun-day but whatever this is just how it worked out.

Stay tuned for the next chapter!