One of my customer called me this past week and told me that his guest Wi-Fi is down and not working. This site does not uses Mist/Juniper equipment so goal of my post is to provide a high level comparison on trouble shooting with Marvis and how fast I am able to find the root cause and resolve an issue. That is all the information I was able to get out of him and his expectation was to resolve the issue quick. I knew there were some recent changes to their network so I had some ideas however I wasn’t certain what exactly was the issue. Let’s assume there were no changes and/or I had no clue of the background for the sake of this example.
It was a typical Network setup with all the main VLANs on the core switch and a routed connection to the router. Second connection to the router was Guest VLAN connected to the router with a L2 connection.
I started my stop watch and started looking. First I looked at their NMS and confirmed that all access points are online, then I saw if there are any other client’s online or not. So far all good. Next I decided to see if there are any clients on the Guest Wi-Fi, there were none. So now I knew that the issue was strictly related to the guest Wi-Fi only. Next step I had to VPN in and then log into their router. Checked the DHCP bindings for the guest clients and there was nothing. Just to make sure all is good in the switch I logged into the switch to confirm everything and all looked good. So now I knew that the issue is with the DHCP. NOTE: This information can be obtained from the user as well but that depends on if the user is willing to work with you or not and other factors. In this case, “Wi-Fi is down” was the only thing user had time for.
So from start to end (i.e; finding the root cause, it took about 8 min approximately).
I decided to replicate the scenario in my Mist Lab and compare the results. I setup a new guest network in my lab and denied access to the DHCP server on that specific VLAN. Next I tried to connect to the new SSID. Disconnected and tried to connect bunch of times to generate some data.
Next step was finding the root cause. I logged into the Mist Dashboard, clicked on Marvis and typed as shown below and there was my root cause with Frame Captures. Total time not even 2 min. Frame capture shows Bootp request but no response back.
I currently do not have Juniper switch, router and/or firewall in my lab, but once I am able to purchase and add those and have an end to end Juniper/Mist network I would love to see how deep Marvis can do in trouble shooting some network related issues that customers normally report.