VMware Networking Community
cnrz
Expert
Expert

NSX Universal Controller Site Redundancy

Controllers in NSX function as the Control Plane for the Logical Switches(LS) and the Logical Routers(DLR). For the LS, MAC address learning and ARP Learning Process, and for the DLR sending of routing updates to routing instances is achieved through Controller. For the Universal LS and Universal DLRs, Universal Controllers provide these functions. Universal Controllers are associated with the Primary NSX Manager on Site A.

In a 3 Site design, losing site A because of a Network Problem how will the Site2 and Site3 be effected? Is there a possibility to stretchor Distribute  the Controllers to 3 Sites,  1 Controller for each(as 3 controller is supported but theoretically it may be an odd number as 9), or recover these controllers on site B if with using Vmotion or SRM? But if they are tied to the Primary NSX Manager, I think not possible to register to Secondary NSX Managers.  If Site2  and 3 cannot reach the  controllers during this time frame the Universal LS and DLRs may have problems with new VM addtions, Vmotion of VMs. What may be recommandations?

http://blogs.vmware.com/networkvirtualization/2016/03/cross-vc-nsx-multi-site-solutions.html#more-27...

Universal_Syncronization_Service.jpg

One workaraound may be to use Multicast Replication Mode, but Multicast service is not supported for many Service providers in an MPLS environment, and it makes the control plane design complicated.

http://chansblog.com/tag/nsx-dlr/

Universal_Controller_DLR.jpg

Best Regards,

Reply
0 Kudos
2 Replies
cnrz
Expert
Expert

The following procedure may be helpful in this rare but important case:(Does not seem to take long time, but may be an automated simlar to SRM use:

http://blogs.vmware.com/vsphere/2014/05/automate-failover-with-srm.html)

Recovery the Management/control plan:

  • Log in to secondary NSX Manager and then Promote Secondary NSX Manager to Primary by: Assign Primary Role.
  • Deploy new Universal Controller Cluster and synchronize all objects
  • Universal CC configuration pushed to ESXi Hosts managed by Secondary
  • Redeploying the UDLR Control VM.

http://www.routetocloud.com/2016/02/nsx-dual-activeactive-datacenters-bcdr/#Complete_Edges_cluster_f...

Universal_Manager_Site_Failure_Recovery.jpg

Reply
0 Kudos
cnrz
Expert
Expert

Similar Threads and Disaster Recovery with NSX and SRM Document is below:

Basic NSX Question in Dual Site Setup https://communities.vmware.com/message/2593791#2593791

Failover between sites with Primary and Secondary NSX Manager Configuration - What about the controllers?https://communities.vmware.com/thread/535356

Disaster Recovery with NSX and SRM https://communities.vmware.com/docs/DOC-31692

This Document "Handling Full Site Failure" heading discuss in detail how to provision secondary site procedure.

Reply
0 Kudos