example

Circonus One Step Install (COSI) examples

Demonstrate running COSI on various operating systems using several different types of automation/orchestration methods. There are two sets of examples basic and advanced. The basic examples demonstrate the COSI process using the default options. The advanced examples illustrate customizing the COSI process with installing a hook, updating NAD plugins, using a custom registration configuration, etc.

OS Options

• CentOS (6.3, 6.6, 6.7, 7.1.1503, 7.2.1511)
• Ubuntu (12.04-trusty, 14.04-precise)
• OmniOS (r151014)

Automation Options (provisioner)

• Manual
• Shell
• Ansible
• Puppet
• coming soon: Chef

Note on manual provisioner:

Simply spins up the VM and does not attempt to run COSI. The main purpose of this provisioner is to provide a set of pre-defined platforms for testing the cut-n-paste COSI command from the API Tokens page.

Since the resulting system will be a VM leveraging the host's network connection, adding the --agent push option to the COSI command line is strongly encouraged.

Prerequisites

• Vagrant (v1.8.1)
• Virtualbox (v5.0.20r106931)
• Optional, Ansible (v2.1.0)

Configuration

1. Copy example-config.yaml to config.yaml.
2. Open config.yaml in an editor.
3. Set up variables specific to the Circonus account:
   1. Log into the Circonus API Tokens page. If there are no API tokens, click New API Token in upper right corner.
   2. Click the (i) next to the token to use and from the COSI command displayed in the overlay:
      1. Copy the --key value, paste into config.yaml as the value for api_key
      2. Copy the --app value, paste into config.yaml as the value for api_app
4. Set provisioner to be one of "ansible", "manual", "puppet", or "shell". The default is "manual". (Note: Ansible must be installed locally if the provisioner is set to "ansible". Puppet will be installed on the VM as it is created.)
5. Enable at least one of the VMs.
6. Save all of the changes to config.yaml.

Using

1. Select either basic or advanced.
2. Change to that directory cd basic or cd advanced.
3. Copy config.yaml created in Configuration section above, cp ../config.yaml . or create a symlink to use the same configuration in both basic and advanced directories ln -s ../config.yaml
4. If necessary, make any changes to config.yaml, e.g. enable/disable VMs, etc.
5. Run vagrant up to start. If more than one VM has been enabled, optionally, run vagrant up <vm name> to start a specific VM.

List available VMs

vagrant status

Log into VM

vagrant ssh or vagrant ssh <vm name>

Halting a VM

This will stop the VM, it can be restarted using vagrant up.

vagrant halt or vagrant halt <vm name>

Destroying a VM

There is a script named destroy.sh in the root example directory. Using this to destroy/remove VMs, whether started in basic or advanced directories, will make cleanup much easier. Although Vagrant has a destroy command, it does not know about the checks, graphs, worksheets, etc. that COSI created in Circonus. The destroy script will run a command (/opt/circonus/cosi/bin/cosi reset --all) on the VM which will remove these artifacts so that they do not have to be manually deleted in the Circonus UI.

../destroy.sh <vm name> note, in this case the specific VM to be destroyed is required.

Example

# after setting up config.yaml ('centos7.2' was only vm enabled)

cd basic
ln -s ../config.yaml

vagrant up
# vagrant brings up the 'centos7.2' vm based upon settings in config.yaml

../destroy.sh centos7.2

# 1. 'centos7.2' is verified to be running.
# 2. cosi reset command is run on the vm to remove all artifacts created in Circonus.
# 3. the vagrant destroy command is run to remove the vm.

Troubleshooting

Details for inspecting/troubleshooting the COSI process can be found on the running VM (vagrant ssh <vm name>) in:

• /opt/circonus/cosi/log/install.log