nsnidankoairvm
Contributor
Contributor

Deploying NSX Controller 6.3.3 - no ip assigned

Jump to solution

I am having trouble deploying first NSX controller - for some reason it fails to assign IP address:

lab.PNG

After timeout controller gets deleted by NSX Manager

2017-11-08 03:51:29.591 GMT  INFO taskScheduler-7 RetryExec:97 - I/O exception (java.net.NoRouteToHostException) caught when processing request to {s}->https://10.10.x.212:443: No route to host (Host unreachable)

Network is fully operational - I was able to access NSX Manager from the test VM on the same network where NSX controller is getting deployed.

Also, I am unable to log into controller via VM console with admin and password I've assigned during deployment.

Can someone point me in the right direction?

Thank you

1 Solution

Accepted Solutions
ShanVMLand
Expert
Expert

I had the same issue, looks like that "root" password expiry issue. The attached PDF explains the procedure to fix the issue.

Thanks,

Shan.

PS: Thanks for your screenshot.

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!

View solution in original post

10 Replies
Sreec
VMware Employee
VMware Employee

No route to host (Host unreachable) basically means there is a problem with DNS/Gateway . I believe the screenshot what you have captured before the controllers getting deleted ?  So double check below points.

1. Ensure that you are connecting to right IP-Pool during deployment and entries are correct .

2. Do check DNS,Gateway connectivity from the Portgroup where Controllers are getting deployed also DNS/NTP  between NSX Manager/vCenter/ESXi hosts

3. Any firewall between Controllers and NSX manager ?

Cheers,
Sree | CKA|CKAD|VCIX-3X| VCAP-4X| VExpert 5x
0 Kudos
nsnidankoairvm
Contributor
Contributor

Everything is configured properly, in fact if i deploy test VM with the same network setting we assign to a controller it can talk to NSX Manager.

I've tried to connect to the controller via VM console in "Deploying" state, however it doesn't accept my credentials - I've tried both root and admin account. It seems that startup script (inside the Controller VM), which should take ovf application properties doesn't assign them correctly.

Is there a way i can access Controller VM, while it is getting deployed?

0 Kudos
Sreec
VMware Employee
VMware Employee

I'm sure you are missing something very basic here .- wrong prefix/gateway/DNS etc are something very common what i have seen for these type of failures. Can you please provide NSX manager IP stack and controller IP pool information ,if that is not possible ,kindly retry the steps and post the NSX manager logs. Also do check if you have enough compute/storage resource on the underlying ESXI host.

Cheers,
Sree | CKA|CKAD|VCIX-3X| VCAP-4X| VExpert 5x
0 Kudos
ShanVMLand
Expert
Expert

I had the same issue, looks like that "root" password expiry issue. The attached PDF explains the procedure to fix the issue.

Thanks,

Shan.

PS: Thanks for your screenshot.

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!

View solution in original post

Sreec
VMware Employee
VMware Employee

Did you managed to find a solution. Please do check VMware Knowledge Base   as well.

Cheers,
Sree | CKA|CKAD|VCIX-3X| VCAP-4X| VExpert 5x
0 Kudos
nsnidankoairvm
Contributor
Contributor

Thank you this was a solution! Also I've upgraded Manager to to 6.3.4 and never had any issues with deploying Controllers - tested multiple times.

0 Kudos
sreeve2929
Contributor
Contributor

I'm getting the same error. (no ip address assigned and the same error/logs in the NSX Manager logs)

I don't have this file on the nsx controller:

/etc/system/network/10-eth0.network

so I'm not sure what to do about that (saw that in the PDF file)

Is this issue specific to 6.3.3?

I have an entry in my dns server for "nsx-controller.<domain>"

pings, dig, nslookup commands all look normal.

Anyone recommend me to upgrade to a later version of NSX?

0 Kudos
nsnidanko88
Contributor
Contributor

yes update to 6.3.5 (or even 6.3.4) and your issue will be resolved.

0 Kudos
sreeve2929
Contributor
Contributor

Problem solved.

Yes, that PDF document was great.

Just statically assigned the address and it came right up.

There is a typo in that document.

4. Type the following command:

vi /etc/system/network/10-eth0.network

that should be "systemd".

0 Kudos
csmith70
Contributor
Contributor

In the NSX-ControllerDepoyment-Issue.pdf file, #4 has a typo in the command it tells you to type (it is correct in the image though).  It should read

vi /etc/systemd/network/10-eth0.network

It appears that even here, spell check removes the d from the systemd directory name.

0 Kudos