Automated ova download and deploy

jaimelm · February 8, 2021, 1:41am

I’ve posted some simple bash scripts for pulling the latest .ova for a given FCOS stream and deploying a node to vSphere based on that image. A little timesaver, particularly for testing.

github.com

JaimeMagiera/fcos-tools/blob/main/EXAMPLES.md

# Examples

Here are examples of using the scripts in this repository to provision and manage FCOS nodes. They utilize the [govc](https://github.com/vmware/govmomi/tree/master/govc) command line tool and assume you've set your govc environment variables for cluster URL, username, password, etc.

## Loaded the latest vSphere template available for a given FCOS stream ##
In this example, the latest release template from the FCOS *next* stream is imported into the vSphere Content Library "Linux ISOs". The script can pull the latest .ova from the stable, testing, and next streams. If the template already exists in the library, the script outputs a message and exits.
``` Bash
#!/bin/bash

manage-release-template.sh --stream "next" --library "Linux ISOs"
```

## Deploying a vm from a vSphere FCOS template with ignition and k-args
In this example, it's assumed that the template "fedora-coreos-33.20201201.3.0-vmware.x86_64" exists the vSphere Content Library "Linux ISOs"

``` Bash
#!/bin/bash

deploy-coreos-node.sh --ova fedora-coreos-33.20201201.3.0-vmware.x86_64 --ignition "fcos-next.ign" --name "fcos-next" --cpu 2 --memory 4000 --disk 100 --folder "/MyCluster/vm/Linux/FCOS/" --library "Linux ISOs" --ipcfg "ip=10.211.2.92::10.211.0.1:255.255.255.0:${VM_NAME}:${IFACE}:off" --boot
```

This file has been truncated. show original

lucab · February 8, 2021, 7:53am

Nice!
From a quick skim it looks like a smart wrapper around the bare govc instructions we have in our docs.
While writing that, did you maybe find something that can be directly added there?

jaimelm · February 10, 2021, 3:56pm

Yes, there are some additions that come to mind.

rugk · April 26, 2021, 11:01pm

If the docs are meant, so do you consider contributing to them, @jaimelm?
Or at least list/draft some ideas of what you would add in a GItHub issue or so.