Cisco has announced that after MSO 3.1, they will no longer provide the .ova files, so you cannot run the MSO as a docker swarm after MSO 3.1. Cisco is REQUIRING that customers must run the MSO from 3.2 on using the NEXUS Dashboard as a hosting platform, then download the MSO Application and install it on the Cisco NEXUS Dashboard. There are also rebranding efforts, so the Cisco NEXUS Dashboard becomes a cohesive platform to run applications. MSO as of 3.2 is referred to as Cisco NEXUS Dashboard Orchestrator and runs as an application on either the physical or virtual Cisco NEXUS Dashboard.

When I discuss the upcoming changes to MSO deployment with customers, most think that they need to buy a physical NEXUS Dashboard cluster to run just MSO. Today that is true, but coming in the next month, the Virtual NEXUS Dashboard(vND) becomes available. The vND is deployed as a cluster of VM's using OVA's and performs running the identical code versions as the physical NEXUS Dashboard. The good news is that the vND is free of charge, so for those looking to migrate their MSO to 3.2, I would advise them to wait until the vND comes out and migrate, which is very straightforward.

So the question I get is, why is Cisco doing this? First and foremost is they are building a unified hosting platform(NEXUS Dashboard) with a correlated database where you can provide automation for site configuration using the NEXUS Dashboard Orchestrator(MSO) and gather telemetry for consumption by day 2 ops tools. To do this, they need to take all of the standalone apps for day 2 operations(MSO, NAE, NIR, and NIA) and put them in a standard application format that can leverage NEXUS Dashboard and its correlated database. Cisco has currently combined the NIR and NIA standalone app functions into NEXUS Insights, and they have ported over MSO 3.2 app to the new format. In future releases, NAE is combined into the NEXUS Insights and become Cisco NEXUS Dashboard Insights.

I recommend that customers facing this migration scenario use this opportunity to look at the NAE and NI day 2 tools. Use the software trials to become proactive in monitoring your network and lower MTTR. The vND is free, and you can install NAE and NI and start monitoring your 9K fabrics, whether they are ACI, VXLAN EVPN, or plain vPC 3 tier. Just be aware that the vND is for smaller fabrics; for larger fabrics, the physical ND is required. Use the two new on-demand labs just released on the WWT Platform for setting up NEXUS Dashboard and NEXUS Insights and NAE with vSimulator. These hands-on labs give you all the use cases for monitoring and troubleshooting your DC networks and become a proactive operator of your data center network. 

Migration from the MSO VM Docker swarm to NEXUS Dashboard is very straightforward. First, make sure you onboard the sites that MSO manages onto the NEXUS Dashboard to migrate to the MSO App on ND. Keep in mind that the naming MUST match the naming in the MSO you are migrating from. Below we see the 3 sites, MSITE400-EAST, MSITE400-WEST, and MSITE400-EQUINIX; in MSO, they must be named the same in ND

Note the names in MSO

They MIST match in NEXUS Dashboard sites.

Next perform a backup on the MSO you want to migrate, and download to your desktop

Shut down your 3 MSO VMs

Log back into NEXUS Dashboard and open up the MSO App

Next go into your MSO App and import the backup file and rollback to the backup 

You will now see that your Sites are being managed by the MSO App on the NEXUS Dashboard. You can makes changes and operate your MSO as before and use the new features in 3.2

The process is very simple and since the vNEXUS Dashboard will be available for free in the next month there should be no issues migrating to the NEXUS Dashboard platform and also explore the day 2 operations tools for free until the grace period is up. Looking back at the NEXUS Dashboard Sites we see that Services used now includes Multi-Site Orchestrator.