This post will be the second post on Nutanix Calm and will build up on my first post which can be found here: http://www.vcoportal.de/2018/09/nutanix-calm-a-deeper-look-from-an-automation-consultant-perspective-part-1/ if you missed it.
In the last post we only created a virtual machine without any customization. In this post we will install additional packages and add an SSH Key for login to the VM.
Let’s go back to our previously created Blueprint. On the left site we have the tab “Package”. Choose this, give the Package a Name and use “Configure install”
What you will see is that beside the LinuxServer Icon a new Icon will open up. Here we choose “+ Task” to create some modifications during the VM build process.
On the left site you must configure some values. You provide a name, choose what you want to do (I want to execute a Shell Script), and take the Credentials you need to access the VM. In the script section you past your script (Linux Bash programming in my case) which you want to execute.
My script for example looks like this:
#!/bin/bash
cat <<'EOF' > /etc/resolv.conf
# Generated by NetworkManager
search vcoportal.de
nameserver 192.168.1.10
nameserver 192.168.1.20
EOF
yum -y update
yum -y install net-tools traceroute bind-utils
First, I declare my script as bash script (line 1). Then I change the DNS Server configuration (line 2 to 7) and update everything on my VM and then install some tools like dig, traceroute and so on for my virtual machine.
When you are finished, save the Blueprint and give it a try for a run.
In the Audit Section of the Application you can take a look at your installation and configuration. As you will see the new generated section InstallSoftware will also be there as point. You can see what the Output of the section is. When you click on the Small window in the market red rectangle you can pop up a bigger window with the output.
On the button you can also have a look at the created script.
In the next step we will insert an SSH Key. This key can be insert from the user wo execute the Blueprint. So, let’s go back to our Blueprint. Here we go the the “Service” Section from our Service. There we insert a Name for the Variable (in my case SSHKey) and a default value. It is important that you activate the “Runtime” checkbox. We want the SSHKey auf our user, not the default and so the user can insert “his” key.
When we are finished with the variable we add a “Task”. This task we use to insert the SSH Key into the authorized_keys field for a specific account (I am lacy so I use the root account. For security reason I would not recommend this in production….)
The script itself is also a Linux Bash script and easy.
#!/bin/bash
mkdir /root/.ssh
touch /root/.ssh/authorized_keys
echo "@@{SSHKey}@@" >> /root/.ssh/authorized_keys
In the first line we declare the script as a Linux Bash script. The second line create the necessary directory and the third line creates the file. In the last line we insert something in the create file……
Did you recommend the “@@{SSHKey}@@ in the last line? What’s that for a strange value?
In Calm variables which we use in script are market with “@@{VALUE}@@” so in our case that means we take the value of the before created Variable and replace the “@@{SSHKey}@@” field with the content.
One note why I used a new task and not placed the script in the first task we already did. From my point of view, I would recommend to create a new task for every goal which I want to reach. I personally thing that debugging is much easier in this case because we can see what went wrong (in which script). The next thing for me is, that even on script maybe doesn’t function like expected we still have the chance to access the VM and fix it during life time.
After we are done with our script, save the Blueprint and Launch the VM. When we are on the Blueprint just open the “Configuration” part and insert the SSH Key which you want to access the VM without password.
When everything works like expected you can login in your VM via SSH. If not, take a look in the Audit section.
As you can see the start with calm is very easy and there are a lot of things we can be done with calm.
During my “play around” with Calm I found some things which must be improved within Calm or limit Calm at the moment:
- Calm is limited to Nutanix, vSphere-Cluster and Cloud Solutions. So, some customers still have to use a second solution for their automation need.
- You cannot export created Blueprint and import them with all needed scripts, variables and images. With our easy Blueprint that is not a big issue but when you create a Blueprint with plenty of variables there is a “room” for mistakes and typos.
- At the moment Calm has some limitations to work with for example ServiceNow and so on.
Some Ideas for Calm for future functions / extension:
- RFE: Blueprint Full Backup with images, scripts, cloudinit etc. à Provide a central repository in PC to make things re-usable: cloudinit, sysprep, scripts, images etc.
- RFE: DNS-Update Feature when VXLAN-DHCP from PE assigns IP-address à Autofill connection URL after application deployment (aka openshift) in the app description
- RFE: Integration of 3rd-Party config-mgmt tools like git, puppet, chef etc. in the GUI (dropdown menu, not via scripts) à Make existing scripts / tools / 3rd-party development re-useable
- RFE: Must-have: Native container integration to push CALM! No customer with an existing automation or configuration solution will switch to CALM without that feature.
- RFE: VM customization options with cloudinit for linux and sysprep for windows delivered as one-click/dropdown options in CALM (forgot that in the call)
If you can life with this limitations Calm is a good choice to start automation in your Nutanix environment.
Steve Jin: VMware VI and vSphere SDK
David Flanagan: JavaScript – The Definitive Guide
).
Marijn Haverbeke: Eloquent JavaScript
Cody Bunch: Automating vSphere: With VMware vCenter Orchestrator