Default Deployment of NSX-T

[MrVmware9423]

Dear team,

As per NSX-T Reference design design guide,
The default (and the simplest option, with the least number of IP addresses consumed) is to
deploy a 3-node cluster without any additional configuration. With this option, each node is
accessible via distinct IP address (or FQDN) and thus multiple endpoints (user and automation
systems) can access each different node to build redundancy and load balancing model for NSXT Manager role. Here the availability is driven via external system choosing different IP address
(FQDN). However, in case of the node failure the system using that node must externally
intervene to point to another available node. For an example, if vRA or API script uses the
FQDN of node A, in case node A fails, there has to be some manual intervention to make the
script continues to work, either you change the FQDN in your API script, or update the FQDN
entry in the DNS. In this configuration, using round robin DNS will result in intermittent
connectivity in a failure scenario. But in terms of topology requirement, as long as there is IP
connectivity between all the nodes, this mode will work.

I'm not able to understand the point marked in BOLD

if we are going to access NSX-T Manager directly from it's FQDN for eg Node A,  if it fails for some reason then how it will make any difference if we  update the FQDN entry in the DNS and can access through round robin DNS.

Could someone please throw more light might be my understanding is not correct.

thanks in advance.

[RickVerstegen]

This is when you are using automation tools that point to a node directly, for example node a. When that node fails you cannot use the automation tool anymore for execute API calls unless you specify it to use one of the remaining nodes that are available, for example node b or node c. This requires manual updating. To access the NSX-T Manager UI, you just go to node b or node c directly when node a is down. Hope this helps.

[MrVmware9423]

thank you sir, for crystal clear explanation.

below statement as per Reference design guide

"if we are going to access NSX-T Manager directly from it's FQDN for eg Node A,  if it fails for some reason then how it will make any difference if we  update the FQDN entry in the DNS and can access through round robin DNS."

I would like to understand, if node A fails then we need to do the changes in the automation tool, so that we can access other available node b or node c, I'm not able to understand why we need to update the FQDN entry in DNS and how we can access through round robin as I'm accessing individual node directly then how it will be considered as round roubing DNS, Please assist on the same.

[RickVerstegen]

In my opinion this indeed not applicable for default deployment setup, unless you have created 3 (node a, node b, node c) A records with different IP addresses with the same FQDN When reaching that FQDN it can still query the failed node and you don't get a response because it's down. So you need to update that record with a (new) node that is online or remove the A record from participating in the cluster in DNS. 

[MrVmware9423]

What i've understood is we have one FQDN with 3 IPaddresses in DNS hence we need to remove the failed node IP address from DNS and post which we would be able to access smoothly, Please correct me If I'm wrong...

Thank you in advance..