I wanted to take the opportunity here to introduce a new set of tools shipped together with the recently released NSO 4.1 release.
It's called nct, NSO cluster tools. The purpose of this toolset is to aid the operational team, managing and running a large scale NSO installation.
The typical tasks that face the operator in a clustered NSO installation
- Upgrading all NSO nodes
- Upgrading NEDs/Applications on all NSO nodes
- Act after HA failovers.
- Regular system management on all NSO nodes, managing e.g users, logs, groups, backups, system settings etc.
There already exists a multitude of great tools to do some of these tasks, one of my favourite tools is Ansible, http://www.ansible.com/ , which we used in CloudVPN.
One shouldn't underestimate how error prone the process of manually reconfiguring a set of nodes is. Actually, come to think of it, this is the very reason why we initially built NSO - to be able to automate configuration changes on a set of routers.
Manually going through a set of NSO nodes and manually performing a simple change, say, changing syslog settings on all hosts is almost bound to fail.
The nct tools are ideal for this task, and it's in many ways superior to ansible for the actual task we have - managing NSO nodes. It also lends itself better to ad hoc tasks than Ansible. Anyways, getting started with nct is easy. It's shipped with NSO and it comes with excellent documentation in the form of man pages.
$ man nct
NCT(1) NCS Manual NCT(1)
nct - a collection of tools that can be used to install and manage NCS
I see two main usage patterns for nct.
- One, doing exactly what we did with Ansible in the cloudVPN (vMS) project. Write explicit - project specific - upgrade scripts. This is obviously on a per project basis.
- Second, as a power tool for the operator running an NSO installation.
So - in your next NSO installation - I recommend that you take a good look at these tools and use them.