We all know and love NetQ – it works hand-in-hand with Linux to accelerate data center operations. Customers love how easy it is to install and operate which makes their lives easier. Also, it can prevent and find issues in a data center by viewing the entire data center as a whole and providing three different types of services:
- Preventative: NetQ allows an engineer to check all data center configurations and state in a few steps from any location in the network. The validation can be done on a virtual network using vagrant with Cumulus VX or if a virtual environment is not available, it can also be used during an change outage window. Since NetQ has built in analyzers of the network as a whole, no scripting is required and the validation is done from one location, rather than hop by hop. It can also shorten outage windows needed for network changes allowing shorter outage windows virtually or during outage windows.
- Proactive: NetQ supplies notifications if something goes wrong in the network by either logging it to a file or integrating with third party applications like Slack, PagerDuty, or Splunk. It can also be filtered to ensure the right team gets the right message.
- Diagnostic: NetQ also provides the ability to troubleshoot your entire data center end to end – from only one location. It also provides any past network state from anytime.
NetQ is supported on both switches and hosts, fully unifying data center visibility. However, some customers come with the question, when should I install NetQ on my linux hosts? While you’ll still receive many benefits without installing NetQ on the hosts, installing it on the hosts will provide an unparalleled visibility into your end-to-end network.
What benefits do I get installing NetQ on a host?
If you are running routing on the host, there are huge benefits to also installing NetQ. With routing on the hosts, the hosts participate in and become part of the data center fabric. Including the hosts, NetQ features such as “netq check bgp” that now becomes integral to analyzing your holistic data center.
If you elect not to run routing on your hosts, NetQ on a host still plays a major function in your day to day data center operations. Running a NetQ agent on a host provides:
- Information on the health and inventory of the host
- Connectivity information and analysis from the host to the network
- Visibility into container activity, connectivity and impact
Information on the health and the inventory of the hosts
When you install the NetQ agent on a host, the NetQ agent is able to communicate information about the hosts to the Telemetry Server. With this, you are able to see things like memory, os version, sensors, and uptime- to name just a few. The example below shows the OS version along with the switches.
Connectivity Information and Analysis from the Host to the Network
When you install NetQ on a host, the host-ToR links are included in the overall MTU check. This way we can be sure the uplink server MTU matches the ToR downlink MTU, reducing the need for fragmentation and other issues in your network. For example, by having NetQ installed on the host with a mismatch, you would see the following when performing a netq check mtu:
Duplicate IP addresses
NetQ can also detect duplicate IP addresses. Since duplicate IP addresses are not only found on switches, why should NetQ be?
You can even see when the change occurred:
Since NetQ listens to Netlink (primarily the interprocess communication between the linux kernel and the userspace), we are informed the minute an interface goes down. Additionally, NetQ checks the end to end link for errors. In the below example, there are autonegotiation mismatches between the leaf switches and servers.
Trace Server to Server
Installing a NetQ agent on the server also allows a holistic view of application connectivity, directly from server to server.
Container Connectivity and Impact
If you are running containers in your network, NetQ also provides huge benefits. With one command, NetQ allows you to see exactly how the different services are connected to your network as well as what it looked like in the past.
If you want to look at the deployment impact if we were to swap out leaf03 from the network at this time:
What if I can install on only a few hosts?
Installing NetQ on only a few hosts adds value, especially if you have containers on your servers with Kubernetes. Since NetQ taps directly into the Kubernetes API, all we need is access to the Kubernetes master nodes to view container deployments, services, pods, etc. Keep in mind that without NetQ on all the hosts, we cannot see all of the impact or connectivity that we see above.
For example, by running NetQ on only the kubernetes master we are able to see the cluster locations of all the PODs. Of course, everything we see here we also see with NetQ agent installed on all hosts.
We are also able to see all the POD locations:
And, of course we can still see the history with NetQ deployed on only the master node(s):
So, as you can see, there is tremendous value in installing the NetQ agent on the hosts. Even more-so if you have containers. But, if you cannot, don’t despair! There is still huge benefits to running NetQ on the switches as before.
Try it out for yourself using Cumulus in the Cloud. After you have spun up the free virtualized environment, use the command “netq ts decommission <server-node> purge” remove NetQ from the host, and see the difference for yourself.