Cannot Form ND Cluster in AWS

[AJ Cruz]

I'll start off with this big caveat:

I modified the Cisco-provided Nexus Dashboard CloudFormation template so that Elastic IPs are not assigned to the Management interfaces. Instead I'm using a DMVPN deployment to connect my VPC to my on-prem lab.

I have deployed the 3 ND nodes using the modified Cisco-Provided CloudFormation template. All 3 nodes deploy no problem.

I added a permit any rule to the SG created by the template so I could do ping testing.

All 3 nodes can ping each other.

When I HTTPS to Node1, supply the mgmt IP and password for node 2 and hit validate I get:

Failed to verify vnode info: exit status 1.

I know the password is correct because I can console in to Node 2 and log in with the recue-user account.

Any ideas what the problem could be?

Thanks!

[AJ Cruz · Accepted solution]

I just noticed this in the deployment guide:

Note	You must enable password-based login on all nodes before proceeding to cluster bootstrap described in the following steps or you will not be able to complete the cluster configuration.

I didn't enable password-based login with: 

acs login-prompt enable

I can do that to Node 1 using the serial console, but Nodes 2 & 3 seem to be stuck in a boot (or service initialization) loop.

I can ping them steady no problem, but when I connect to console I get this over and over:

system initialized successfully
Please wait for system to boot : [##########------------------] 29%
CentOS Linux 7 (Core)
Kernel 4.14.241stock-2 on an x86_64

system initialized successfully
Please wait for system to boot : [##########------------------] 29%
CentOS Linux 7 (Core)
Kernel 4.14.241stock-2 on an x86_64

and I can't get a login prompt.

I've rebooted, destroyed and re-created the stack. For some reason I cannot console in to Nodes 2 & 3, only Node 1.

UPDATE:

I can't connect to console of 2 & 3, but I can connect via SSH using the certificate.

After enabling the login password the cluster is forming.